| 1 | * Summary: interface for the encoding conversion functions
|
|---|
| 2 | * Description: interface for the encoding conversion functions needed for
|
|---|
| 3 | * XML basic encoding and iconv() support.
|
|---|
| 4 | *
|
|---|
| 5 | * Related specs are
|
|---|
| 6 | * rfc2044 (UTF-8 and UTF-16) F. Yergeau Alis Technologies
|
|---|
| 7 | * [ISO-10646] UTF-8 and UTF-16 in Annexes
|
|---|
| 8 | * [ISO-8859-1] ISO Latin-1 characters codes.
|
|---|
| 9 | * [UNICODE] The Unicode Consortium, "The Unicode Standard --
|
|---|
| 10 | * Worldwide Character Encoding -- Version 1.0", Addison-
|
|---|
| 11 | * Wesley, Volume 1, 1991, Volume 2, 1992. UTF-8 is
|
|---|
| 12 | * described in Unicode Technical Report #4.
|
|---|
| 13 | * [US-ASCII] Coded Character Set--7-bit American Standard Code for
|
|---|
| 14 | * Information Interchange, ANSI X3.4-1986.
|
|---|
| 15 | *
|
|---|
| 16 | * Copy: See Copyright for the status of this software.
|
|---|
| 17 | *
|
|---|
| 18 | * Author: Patrick Monnerat <[email protected]>, DATASPHERE S.A.
|
|---|
| 19 |
|
|---|
| 20 | /if not defined(XML_CHAR_ENCODING_H__)
|
|---|
| 21 | /define XML_CHAR_ENCODING_H__
|
|---|
| 22 |
|
|---|
| 23 | /include "libxmlrpg/xmlversion"
|
|---|
| 24 | /include "libxmlrpg/xmlTypesC"
|
|---|
| 25 |
|
|---|
| 26 | * xmlCharEncoding:
|
|---|
| 27 | *
|
|---|
| 28 | * Predefined values for some standard encodings.
|
|---|
| 29 | * Libxml does not do beforehand translation on UTF8 and ISOLatinX.
|
|---|
| 30 | * It also supports ASCII, ISO-8859-1, and UTF16 (LE and BE) by default.
|
|---|
| 31 | *
|
|---|
| 32 | * Anything else would have to be translated to UTF8 before being
|
|---|
| 33 | * given to the parser itself. The BOM for UTF16 and the encoding
|
|---|
| 34 | * declaration are looked at and a converter is looked for at that
|
|---|
| 35 | * point. If not found the parser stops here as asked by the XML REC. A
|
|---|
| 36 | * converter can be registered by the user
|
|---|
| 37 | * xmlRegisterCharEncodingHandler but the current form doesn't allow
|
|---|
| 38 | * stateful transcoding (a serious problem agreed !). If iconv has been
|
|---|
| 39 | * found it will be used automatically and allow stateful transcoding,
|
|---|
| 40 | * the simplest is then to be sure to enable iconv and to provide iconv
|
|---|
| 41 | * libs for the encoding support needed.
|
|---|
| 42 | *
|
|---|
| 43 | * Note that the generic "UTF-16" is not a predefined value. Instead, only
|
|---|
| 44 | * the specific UTF-16LE and UTF-16BE are present.
|
|---|
| 45 |
|
|---|
| 46 | d xmlCharEncoding...
|
|---|
| 47 | d s based(######typedef######)
|
|---|
| 48 | d like(xmlCenum)
|
|---|
| 49 | d XML_CHAR_ENCODING_ERROR... No encoding detected
|
|---|
| 50 | d c -1
|
|---|
| 51 | d XML_CHAR_ENCODING_NONE... No encoding detected
|
|---|
| 52 | d c 0
|
|---|
| 53 | d XML_CHAR_ENCODING_UTF8... UTF-8
|
|---|
| 54 | d c 1
|
|---|
| 55 | d XML_CHAR_ENCODING_UTF16LE... UTF-16 little endian
|
|---|
| 56 | d c 2
|
|---|
| 57 | d XML_CHAR_ENCODING_UTF16BE... UTF-16 big endian
|
|---|
| 58 | d c 3
|
|---|
| 59 | d XML_CHAR_ENCODING_UCS4LE... UCS-4 little endian
|
|---|
| 60 | d c 4
|
|---|
| 61 | d XML_CHAR_ENCODING_UCS4BE... UCS-4 big endian
|
|---|
| 62 | d c 5
|
|---|
| 63 | d XML_CHAR_ENCODING_EBCDIC... EBCDIC uh!
|
|---|
| 64 | d c 6
|
|---|
| 65 | d XML_CHAR_ENCODING_UCS4_2143... UCS-4 unusual order
|
|---|
| 66 | d c 7
|
|---|
| 67 | d XML_CHAR_ENCODING_UCS4_3412... UCS-4 unusual order
|
|---|
| 68 | d c 8
|
|---|
| 69 | d XML_CHAR_ENCODING_UCS2... UCS-2
|
|---|
| 70 | d c 9
|
|---|
| 71 | d XML_CHAR_ENCODING_8859_1... ISO-8859-1 ISOLatin1
|
|---|
| 72 | d c 10
|
|---|
| 73 | d XML_CHAR_ENCODING_8859_2... ISO-8859-2 ISOLatin2
|
|---|
| 74 | d c 11
|
|---|
| 75 | d XML_CHAR_ENCODING_8859_3... ISO-8859-3
|
|---|
| 76 | d c 12
|
|---|
| 77 | d XML_CHAR_ENCODING_8859_4... ISO-8859-4
|
|---|
| 78 | d c 13
|
|---|
| 79 | d XML_CHAR_ENCODING_8859_5... ISO-8859-5
|
|---|
| 80 | d c 14
|
|---|
| 81 | d XML_CHAR_ENCODING_8859_6... ISO-8859-6
|
|---|
| 82 | d c 15
|
|---|
| 83 | d XML_CHAR_ENCODING_8859_7... ISO-8859-7
|
|---|
| 84 | d c 16
|
|---|
| 85 | d XML_CHAR_ENCODING_8859_8... ISO-8859-8
|
|---|
| 86 | d c 17
|
|---|
| 87 | d XML_CHAR_ENCODING_8859_9... ISO-8859-9
|
|---|
| 88 | d c 18
|
|---|
| 89 | d XML_CHAR_ENCODING_2022_JP... ISO-2022-JP
|
|---|
| 90 | d c 19
|
|---|
| 91 | d XML_CHAR_ENCODING_SHIFT_JIS... Shift_JIS
|
|---|
| 92 | d c 20
|
|---|
| 93 | d XML_CHAR_ENCODING_EUC_JP... EUC-JP
|
|---|
| 94 | d c 21
|
|---|
| 95 | d XML_CHAR_ENCODING_ASCII... Pure ASCII
|
|---|
| 96 | d c 22
|
|---|
| 97 |
|
|---|
| 98 | * xmlCharEncodingInputFunc:
|
|---|
| 99 | * @out: a pointer to an array of bytes to store the UTF-8 result
|
|---|
| 100 | * @outlen: the length of @out
|
|---|
| 101 | * @in: a pointer to an array of chars in the original encoding
|
|---|
| 102 | * @inlen: the length of @in
|
|---|
| 103 | *
|
|---|
| 104 | * Take a block of chars in the original encoding and try to convert
|
|---|
| 105 | * it to an UTF-8 block of chars out.
|
|---|
| 106 | *
|
|---|
| 107 | * Returns the number of bytes written, -1 if lack of space, or -2
|
|---|
| 108 | * if the transcoding failed.
|
|---|
| 109 | * The value of @inlen after return is the number of octets consumed
|
|---|
| 110 | * if the return value is positive, else unpredictiable.
|
|---|
| 111 | * The value of @outlen after return is the number of octets consumed.
|
|---|
| 112 |
|
|---|
| 113 | d xmlCharEncodingInputFunc...
|
|---|
| 114 | d s * based(######typedef######)
|
|---|
| 115 | d procptr
|
|---|
| 116 |
|
|---|
| 117 | * xmlCharEncodingOutputFunc:
|
|---|
| 118 | * @out: a pointer to an array of bytes to store the result
|
|---|
| 119 | * @outlen: the length of @out
|
|---|
| 120 | * @in: a pointer to an array of UTF-8 chars
|
|---|
| 121 | * @inlen: the length of @in
|
|---|
| 122 | *
|
|---|
| 123 | * Take a block of UTF-8 chars in and try to convert it to another
|
|---|
| 124 | * encoding.
|
|---|
| 125 | * Note: a first call designed to produce heading info is called with
|
|---|
| 126 | * in = NULL. If stateful this should also initialize the encoder state.
|
|---|
| 127 | *
|
|---|
| 128 | * Returns the number of bytes written, -1 if lack of space, or -2
|
|---|
| 129 | * if the transcoding failed.
|
|---|
| 130 | * The value of @inlen after return is the number of octets consumed
|
|---|
| 131 | * if the return value is positive, else unpredictiable.
|
|---|
| 132 | * The value of @outlen after return is the number of octets produced.
|
|---|
| 133 |
|
|---|
| 134 | d xmlCharEncodingOutputFunc...
|
|---|
| 135 | d s * based(######typedef######)
|
|---|
| 136 | d procptr
|
|---|
| 137 |
|
|---|
| 138 | * Block defining the handlers for non UTF-8 encodings.
|
|---|
| 139 | * If iconv is supported, there are two extra fields.
|
|---|
| 140 |
|
|---|
| 141 | /if defined(LIBXML_ICU_ENABLED)
|
|---|
| 142 | d uconv_t ds based(######typedef######)
|
|---|
| 143 | d align qualified
|
|---|
| 144 | d uconv * UConverter *
|
|---|
| 145 | d utf8 * UConverter *
|
|---|
| 146 | /endif
|
|---|
| 147 |
|
|---|
| 148 | d xmlCharEncodingHandlerPtr...
|
|---|
| 149 | d s * based(######typedef######)
|
|---|
| 150 |
|
|---|
| 151 | d xmlCharEncodingHandler...
|
|---|
| 152 | d ds based(xmlCharEncodingHandlerPtr)
|
|---|
| 153 | d align qualified
|
|---|
| 154 | d name * char *
|
|---|
| 155 | d input like(xmlCharEncodingInputFunc)
|
|---|
| 156 | d output like(xmlCharEncodingOutputFunc)
|
|---|
| 157 | *
|
|---|
| 158 | /if defined(LIBXML_ICONV_ENABLED)
|
|---|
| 159 | d iconv_in * iconv_t
|
|---|
| 160 | d iconv_out * iconv_t
|
|---|
| 161 | /endif LIBXML_ICONV_ENABLED
|
|---|
| 162 | *
|
|---|
| 163 | /if defined(LIBXML_ICU_ENABLED)
|
|---|
| 164 | d uconv_in * uconv_t *
|
|---|
| 165 | d uconv_out * uconv_t *
|
|---|
| 166 | /endif LIBXML_ICU_ENABLED
|
|---|
| 167 |
|
|---|
| 168 | /include "libxmlrpg/tree"
|
|---|
| 169 |
|
|---|
| 170 | * Interfaces for encoding handlers.
|
|---|
| 171 |
|
|---|
| 172 | d xmlInitCharEncodingHandlers...
|
|---|
| 173 | d pr extproc(
|
|---|
| 174 | d 'xmlInitCharEncodingHandlers')
|
|---|
| 175 |
|
|---|
| 176 | d xmlCleanupCharEncodingHandlers...
|
|---|
| 177 | d pr extproc(
|
|---|
| 178 | d 'xmlCleanupCharEncodingHandlers')
|
|---|
| 179 |
|
|---|
| 180 | d xmlRegisterCharEncodingHandler...
|
|---|
| 181 | d pr extproc(
|
|---|
| 182 | d 'xmlRegisterCharEncodingHandler')
|
|---|
| 183 | d handler value like(xmlCharEncodingHandlerPtr)
|
|---|
| 184 |
|
|---|
| 185 | d xmlGetCharEncodingHandler...
|
|---|
| 186 | d pr extproc('xmlGetCharEncodingHandler')
|
|---|
| 187 | d like(xmlCharEncodingHandlerPtr)
|
|---|
| 188 | d enc value like(xmlCharEncoding)
|
|---|
| 189 |
|
|---|
| 190 | d xmlFindCharEncodingHandler...
|
|---|
| 191 | d pr extproc('xmlFindCharEncodingHandler')
|
|---|
| 192 | d like(xmlCharEncodingHandlerPtr)
|
|---|
| 193 | d name * value options(*string) const char *
|
|---|
| 194 |
|
|---|
| 195 | d xmlNewCharEncodingHandler...
|
|---|
| 196 | d pr extproc('xmlNewCharEncodingHandler')
|
|---|
| 197 | d like(xmlCharEncodingHandlerPtr)
|
|---|
| 198 | d name * value options(*string) const char *
|
|---|
| 199 | d input value like(xmlCharEncodingInputFunc)
|
|---|
| 200 | d output value like(xmlCharEncodingOutputFunc)
|
|---|
| 201 |
|
|---|
| 202 | * Interfaces for encoding names and aliases.
|
|---|
| 203 |
|
|---|
| 204 | d xmlAddEncodingAlias...
|
|---|
| 205 | d pr extproc('xmlAddEncodingAlias')
|
|---|
| 206 | d like(xmlCint)
|
|---|
| 207 | d name * value options(*string) const char *
|
|---|
| 208 | d alias * value options(*string) const char *
|
|---|
| 209 |
|
|---|
| 210 | d xmlDelEncodingAlias...
|
|---|
| 211 | d pr extproc('xmlDelEncodingAlias')
|
|---|
| 212 | d like(xmlCint)
|
|---|
| 213 | d alias * value options(*string) const char *
|
|---|
| 214 |
|
|---|
| 215 | d xmlGetEncodingAlias...
|
|---|
| 216 | d pr * extproc('xmlGetEncodingAlias') const char *
|
|---|
| 217 | d alias * value options(*string) const char *
|
|---|
| 218 |
|
|---|
| 219 | d xmlCleanupEncodingAliases...
|
|---|
| 220 | d pr extproc('xmlCleanupEncodingAliases')
|
|---|
| 221 |
|
|---|
| 222 | d xmlParseCharEncoding...
|
|---|
| 223 | d pr extproc('xmlParseCharEncoding')
|
|---|
| 224 | d like(xmlCharEncoding)
|
|---|
| 225 | d name * value options(*string) const char *
|
|---|
| 226 |
|
|---|
| 227 | d xmlGetCharEncodingName...
|
|---|
| 228 | d pr * extproc('xmlGetCharEncodingName') const char *
|
|---|
| 229 | d enc value like(xmlCharEncoding)
|
|---|
| 230 |
|
|---|
| 231 | * Interfaces directly used by the parsers.
|
|---|
| 232 |
|
|---|
| 233 | d xmlDetectCharEncoding...
|
|---|
| 234 | d pr extproc('xmlDetectCharEncoding')
|
|---|
| 235 | d like(xmlCharEncoding)
|
|---|
| 236 | d in * value options(*string) const unsigned char*
|
|---|
| 237 | d len value like(xmlCint)
|
|---|
| 238 |
|
|---|
| 239 | d xmlCharEncOutFunc...
|
|---|
| 240 | d pr extproc('xmlCharEncOutFunc')
|
|---|
| 241 | d like(xmlCint)
|
|---|
| 242 | d handler likeds(xmlCharEncodingHandler)
|
|---|
| 243 | d out value like(xmlBufferPtr)
|
|---|
| 244 | d in value like(xmlBufferPtr)
|
|---|
| 245 |
|
|---|
| 246 | d xmlCharEncInFunc...
|
|---|
| 247 | d pr extproc('xmlCharEncInFunc')
|
|---|
| 248 | d like(xmlCint)
|
|---|
| 249 | d handler likeds(xmlCharEncodingHandler)
|
|---|
| 250 | d out value like(xmlBufferPtr)
|
|---|
| 251 | d in value like(xmlBufferPtr)
|
|---|
| 252 |
|
|---|
| 253 | d xmlCharEncFirstLine...
|
|---|
| 254 | d pr extproc('xmlCharEncFirstLine')
|
|---|
| 255 | d like(xmlCint)
|
|---|
| 256 | d handler likeds(xmlCharEncodingHandler)
|
|---|
| 257 | d out value like(xmlBufferPtr)
|
|---|
| 258 | d in value like(xmlBufferPtr)
|
|---|
| 259 |
|
|---|
| 260 | d xmlCharEncCloseFunc...
|
|---|
| 261 | d pr extproc('xmlCharEncCloseFunc')
|
|---|
| 262 | d like(xmlCint)
|
|---|
| 263 | d handler likeds(xmlCharEncodingHandler)
|
|---|
| 264 |
|
|---|
| 265 | * Export a few useful functions
|
|---|
| 266 |
|
|---|
| 267 | /if defined(LIBXML_OUTPUT_ENABLED)
|
|---|
| 268 | d UTF8Toisolat1 pr extproc('UTF8Toisolat1')
|
|---|
| 269 | d like(xmlCint)
|
|---|
| 270 | d out 65535 options(*varsize) unsigned char (*)
|
|---|
| 271 | d outlen like(xmlCint)
|
|---|
| 272 | d in * value options(*string) const unsigned char*
|
|---|
| 273 | d inlen like(xmlCint)
|
|---|
| 274 |
|
|---|
| 275 | /endif LIBXML_OUTPUT_ENABLD
|
|---|
| 276 |
|
|---|
| 277 | d isolat1ToUTF8 pr extproc('isolat1ToUTF8')
|
|---|
| 278 | d like(xmlCint)
|
|---|
| 279 | d out 65535 options(*varsize) unsigned char (*)
|
|---|
| 280 | d outlen like(xmlCint)
|
|---|
| 281 | d in * value options(*string) const unsigned char*
|
|---|
| 282 | d inlen like(xmlCint)
|
|---|
| 283 |
|
|---|
| 284 | /endif XML_CHAR_ENCODING_H
|
|---|
