Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 1 | /* |
| 2 | * encoding.h : interface for the encoding conversion functions needed for |
| 3 | * XML |
| 4 | * |
| 5 | * Related specs: |
| 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 | * |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 16 | * See Copyright for the status of this software. |
| 17 | * |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 18 | * Daniel.Veillard@w3.org |
| 19 | */ |
| 20 | |
Daniel Veillard | 27d8874 | 1999-05-29 11:51:49 +0000 | [diff] [blame] | 21 | #ifndef __XML_CHAR_ENCODING_H__ |
| 22 | #define __XML_CHAR_ENCODING_H__ |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 23 | |
Daniel Veillard | 361d845 | 2000-04-03 19:48:13 +0000 | [diff] [blame] | 24 | #include <libxml/xmlversion.h> |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 25 | #ifdef LIBXML_ICONV_ENABLED |
| 26 | #include <iconv.h> |
| 27 | #endif |
| 28 | #include <libxml/tree.h> |
| 29 | |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 30 | #ifdef __cplusplus |
| 31 | extern "C" { |
| 32 | #endif |
| 33 | |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 34 | /** |
| 35 | * Predefined values for some standard encodings |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 36 | * Libxml don't do beforehand translation on UTF8, ISOLatinX |
| 37 | * It also support UTF16 (LE and BE) by default. |
| 38 | * |
| 39 | * Anything else would have to be translated to UTF8 before being |
| 40 | * given to the parser itself. The BOM for UTF16 and the encoding |
| 41 | * declaration are looked at and a converter is looked for at that |
| 42 | * point. If not found the parser stops here as asked by the XML REC |
| 43 | * Converter can be registered by the user using xmlRegisterCharEncodingHandler |
| 44 | * but the currentl form doesn't allow stateful transcoding (a serious |
| 45 | * problem agreed !). If iconv has been found it will be used |
| 46 | * automatically and allow stateful transcoding, the simplest is then |
| 47 | * to be sure to enable icon and to provide iconv libs for the encoding |
| 48 | * support needed. |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 49 | */ |
Daniel Veillard | 27d8874 | 1999-05-29 11:51:49 +0000 | [diff] [blame] | 50 | typedef enum { |
| 51 | XML_CHAR_ENCODING_ERROR= -1, /* No char encoding detected */ |
| 52 | XML_CHAR_ENCODING_NONE= 0, /* No char encoding detected */ |
| 53 | XML_CHAR_ENCODING_UTF8= 1, /* UTF-8 */ |
| 54 | XML_CHAR_ENCODING_UTF16LE= 2, /* UTF-16 little endian */ |
| 55 | XML_CHAR_ENCODING_UTF16BE= 3, /* UTF-16 big endian */ |
| 56 | XML_CHAR_ENCODING_UCS4LE= 4, /* UCS-4 little endian */ |
| 57 | XML_CHAR_ENCODING_UCS4BE= 5, /* UCS-4 big endian */ |
| 58 | XML_CHAR_ENCODING_EBCDIC= 6, /* EBCDIC uh! */ |
| 59 | XML_CHAR_ENCODING_UCS4_2143=7, /* UCS-4 unusual ordering */ |
| 60 | XML_CHAR_ENCODING_UCS4_3412=8, /* UCS-4 unusual ordering */ |
| 61 | XML_CHAR_ENCODING_UCS2= 9, /* UCS-2 */ |
| 62 | XML_CHAR_ENCODING_8859_1= 10,/* ISO-8859-1 ISO Latin 1 */ |
| 63 | XML_CHAR_ENCODING_8859_2= 11,/* ISO-8859-2 ISO Latin 2 */ |
| 64 | XML_CHAR_ENCODING_8859_3= 12,/* ISO-8859-3 */ |
| 65 | XML_CHAR_ENCODING_8859_4= 13,/* ISO-8859-4 */ |
| 66 | XML_CHAR_ENCODING_8859_5= 14,/* ISO-8859-5 */ |
| 67 | XML_CHAR_ENCODING_8859_6= 15,/* ISO-8859-6 */ |
| 68 | XML_CHAR_ENCODING_8859_7= 16,/* ISO-8859-7 */ |
| 69 | XML_CHAR_ENCODING_8859_8= 17,/* ISO-8859-8 */ |
| 70 | XML_CHAR_ENCODING_8859_9= 18,/* ISO-8859-9 */ |
| 71 | XML_CHAR_ENCODING_2022_JP= 19,/* ISO-2022-JP */ |
| 72 | XML_CHAR_ENCODING_SHIFT_JIS=20,/* Shift_JIS */ |
Daniel Veillard | 32bc74e | 2000-07-14 14:49:25 +0000 | [diff] [blame] | 73 | XML_CHAR_ENCODING_EUC_JP= 21,/* EUC-JP */ |
| 74 | XML_CHAR_ENCODING_ASCII= 22 /* pure ASCII */ |
Daniel Veillard | 27d8874 | 1999-05-29 11:51:49 +0000 | [diff] [blame] | 75 | } xmlCharEncoding; |
| 76 | |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 77 | /** |
| 78 | * xmlCharEncodingInputFunc: |
| 79 | * @out: a pointer ot an array of bytes to store the UTF-8 result |
| 80 | * @outlen: the lenght of @out |
| 81 | * @in: a pointer ot an array of chars in the original encoding |
| 82 | * @inlen: the lenght of @in |
| 83 | * |
| 84 | * Take a block of chars in the original encoding and try to convert |
| 85 | * it to an UTF-8 block of chars out. |
| 86 | * |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 87 | * Returns the number of byte written, or -1 by lack of space, or -2 |
| 88 | * if the transcoding failed. |
| 89 | * The value of @inlen after return is the number of octets consumed |
| 90 | * as the return value is positive, else unpredictiable. |
| 91 | * The value of @outlen after return is the number of ocetes consumed. |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 92 | */ |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 93 | typedef int (* xmlCharEncodingInputFunc)(unsigned char* out, int *outlen, |
Daniel Veillard | cf46199 | 2000-03-14 18:30:20 +0000 | [diff] [blame] | 94 | const unsigned char* in, int *inlen); |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 95 | |
| 96 | |
| 97 | /** |
Daniel Veillard | cf46199 | 2000-03-14 18:30:20 +0000 | [diff] [blame] | 98 | * xmlCharEncodingOutputFunc: |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 99 | * @out: a pointer ot an array of bytes to store the result |
| 100 | * @outlen: the lenght of @out |
| 101 | * @in: a pointer ot an array of UTF-8 chars |
| 102 | * @inlen: the lenght of @in |
| 103 | * |
| 104 | * Take a block of UTF-8 chars in and try to convert it to an other |
| 105 | * encoding. |
Daniel Veillard | be80396 | 2000-06-28 23:40:59 +0000 | [diff] [blame] | 106 | * Note: a first call designed to produce heading info is called with |
| 107 | * in = NULL. If stateful this should also initialize the encoder state |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 108 | * |
| 109 | * Returns the number of byte written, or -1 by lack of space, or -2 |
| 110 | * if the transcoding failed. |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 111 | * The value of @inlen after return is the number of octets consumed |
| 112 | * as the return value is positive, else unpredictiable. |
| 113 | * The value of @outlen after return is the number of ocetes consumed. |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 114 | */ |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 115 | typedef int (* xmlCharEncodingOutputFunc)(unsigned char* out, int *outlen, |
Daniel Veillard | cf46199 | 2000-03-14 18:30:20 +0000 | [diff] [blame] | 116 | const unsigned char* in, int *inlen); |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 117 | |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 118 | |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 119 | /* |
| 120 | * Block defining the handlers for non UTF-8 encodings. |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 121 | * If iconv is supported, there is two extra fields |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 122 | */ |
| 123 | |
Daniel Veillard | 71b656e | 2000-01-05 14:46:17 +0000 | [diff] [blame] | 124 | typedef struct _xmlCharEncodingHandler xmlCharEncodingHandler; |
| 125 | typedef xmlCharEncodingHandler *xmlCharEncodingHandlerPtr; |
| 126 | struct _xmlCharEncodingHandler { |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 127 | char *name; |
| 128 | xmlCharEncodingInputFunc input; |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 129 | xmlCharEncodingOutputFunc output; |
| 130 | #ifdef LIBXML_ICONV_ENABLED |
| 131 | iconv_t iconv_in; |
| 132 | iconv_t iconv_out; |
| 133 | #endif /* LIBXML_ICONV_ENABLED */ |
Daniel Veillard | 71b656e | 2000-01-05 14:46:17 +0000 | [diff] [blame] | 134 | }; |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 135 | |
Daniel Veillard | a819dac | 1999-11-24 18:04:22 +0000 | [diff] [blame] | 136 | void xmlInitCharEncodingHandlers (void); |
| 137 | void xmlCleanupCharEncodingHandlers (void); |
| 138 | void xmlRegisterCharEncodingHandler (xmlCharEncodingHandlerPtr handler); |
Daniel Veillard | cf46199 | 2000-03-14 18:30:20 +0000 | [diff] [blame] | 139 | xmlCharEncoding xmlDetectCharEncoding (const unsigned char* in, |
| 140 | int len); |
Daniel Veillard | a819dac | 1999-11-24 18:04:22 +0000 | [diff] [blame] | 141 | xmlCharEncoding xmlParseCharEncoding (const char* name); |
Daniel Veillard | be80396 | 2000-06-28 23:40:59 +0000 | [diff] [blame] | 142 | const char* xmlGetCharEncodingName (xmlCharEncoding enc); |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 143 | xmlCharEncodingHandlerPtr xmlGetCharEncodingHandler(xmlCharEncoding enc); |
| 144 | xmlCharEncodingHandlerPtr xmlFindCharEncodingHandler(const char *name); |
Daniel Veillard | cf46199 | 2000-03-14 18:30:20 +0000 | [diff] [blame] | 145 | int xmlCheckUTF8 (const unsigned char *utf); |
Daniel Veillard | 14fff06 | 1999-06-22 21:49:07 +0000 | [diff] [blame] | 146 | |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 147 | int xmlCharEncOutFunc (xmlCharEncodingHandler *handler, |
| 148 | xmlBufferPtr out, |
| 149 | xmlBufferPtr in); |
| 150 | |
| 151 | int xmlCharEncInFunc (xmlCharEncodingHandler *handler, |
| 152 | xmlBufferPtr out, |
| 153 | xmlBufferPtr in); |
Daniel Veillard | be80396 | 2000-06-28 23:40:59 +0000 | [diff] [blame] | 154 | int xmlCharEncFirstLine (xmlCharEncodingHandler *handler, |
| 155 | xmlBufferPtr out, |
| 156 | xmlBufferPtr in); |
Daniel Veillard | 496a1cf | 2000-05-03 14:20:55 +0000 | [diff] [blame] | 157 | int xmlCharEncCloseFunc (xmlCharEncodingHandler *handler); |
Daniel Veillard | 27d8874 | 1999-05-29 11:51:49 +0000 | [diff] [blame] | 158 | |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 159 | #ifdef __cplusplus |
| 160 | } |
| 161 | #endif |
| 162 | |
Daniel Veillard | 27d8874 | 1999-05-29 11:51:49 +0000 | [diff] [blame] | 163 | #endif /* __XML_CHAR_ENCODING_H__ */ |