Home | History | Annotate | Download | only in libxml
      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 #ifdef LIBXML_ICU_ENABLED
     31 #include <unicode/ucnv.h>
     32 #endif
     33 #ifdef __cplusplus
     34 extern "C" {
     35 #endif
     36 
     37 /*
     38  * xmlCharEncoding:
     39  *
     40  * Predefined values for some standard encodings.
     41  * Libxml does not do beforehand translation on UTF8 and ISOLatinX.
     42  * It also supports ASCII, ISO-8859-1, and UTF16 (LE and BE) by default.
     43  *
     44  * Anything else would have to be translated to UTF8 before being
     45  * given to the parser itself. The BOM for UTF16 and the encoding
     46  * declaration are looked at and a converter is looked for at that
     47  * point. If not found the parser stops here as asked by the XML REC. A
     48  * converter can be registered by the user using xmlRegisterCharEncodingHandler
     49  * but the current form doesn't allow stateful transcoding (a serious
     50  * problem agreed !). If iconv has been found it will be used
     51  * automatically and allow stateful transcoding, the simplest is then
     52  * to be sure to enable iconv and to provide iconv libs for the encoding
     53  * support needed.
     54  *
     55  * Note that the generic "UTF-16" is not a predefined value.  Instead, only
     56  * the specific UTF-16LE and UTF-16BE are present.
     57  */
     58 typedef enum {
     59     XML_CHAR_ENCODING_ERROR=   -1, /* No char encoding detected */
     60     XML_CHAR_ENCODING_NONE=	0, /* No char encoding detected */
     61     XML_CHAR_ENCODING_UTF8=	1, /* UTF-8 */
     62     XML_CHAR_ENCODING_UTF16LE=	2, /* UTF-16 little endian */
     63     XML_CHAR_ENCODING_UTF16BE=	3, /* UTF-16 big endian */
     64     XML_CHAR_ENCODING_UCS4LE=	4, /* UCS-4 little endian */
     65     XML_CHAR_ENCODING_UCS4BE=	5, /* UCS-4 big endian */
     66     XML_CHAR_ENCODING_EBCDIC=	6, /* EBCDIC uh! */
     67     XML_CHAR_ENCODING_UCS4_2143=7, /* UCS-4 unusual ordering */
     68     XML_CHAR_ENCODING_UCS4_3412=8, /* UCS-4 unusual ordering */
     69     XML_CHAR_ENCODING_UCS2=	9, /* UCS-2 */
     70     XML_CHAR_ENCODING_8859_1=	10,/* ISO-8859-1 ISO Latin 1 */
     71     XML_CHAR_ENCODING_8859_2=	11,/* ISO-8859-2 ISO Latin 2 */
     72     XML_CHAR_ENCODING_8859_3=	12,/* ISO-8859-3 */
     73     XML_CHAR_ENCODING_8859_4=	13,/* ISO-8859-4 */
     74     XML_CHAR_ENCODING_8859_5=	14,/* ISO-8859-5 */
     75     XML_CHAR_ENCODING_8859_6=	15,/* ISO-8859-6 */
     76     XML_CHAR_ENCODING_8859_7=	16,/* ISO-8859-7 */
     77     XML_CHAR_ENCODING_8859_8=	17,/* ISO-8859-8 */
     78     XML_CHAR_ENCODING_8859_9=	18,/* ISO-8859-9 */
     79     XML_CHAR_ENCODING_2022_JP=  19,/* ISO-2022-JP */
     80     XML_CHAR_ENCODING_SHIFT_JIS=20,/* Shift_JIS */
     81     XML_CHAR_ENCODING_EUC_JP=   21,/* EUC-JP */
     82     XML_CHAR_ENCODING_ASCII=    22 /* pure ASCII */
     83 } xmlCharEncoding;
     84 
     85 /**
     86  * xmlCharEncodingInputFunc:
     87  * @out:  a pointer to an array of bytes to store the UTF-8 result
     88  * @outlen:  the length of @out
     89  * @in:  a pointer to an array of chars in the original encoding
     90  * @inlen:  the length of @in
     91  *
     92  * Take a block of chars in the original encoding and try to convert
     93  * it to an UTF-8 block of chars out.
     94  *
     95  * Returns the number of bytes written, -1 if lack of space, or -2
     96  *     if the transcoding failed.
     97  * The value of @inlen after return is the number of octets consumed
     98  *     if the return value is positive, else unpredictiable.
     99  * The value of @outlen after return is the number of octets consumed.
    100  */
    101 typedef int (* xmlCharEncodingInputFunc)(unsigned char *out, int *outlen,
    102                                          const unsigned char *in, int *inlen);
    103 
    104 
    105 /**
    106  * xmlCharEncodingOutputFunc:
    107  * @out:  a pointer to an array of bytes to store the result
    108  * @outlen:  the length of @out
    109  * @in:  a pointer to an array of UTF-8 chars
    110  * @inlen:  the length of @in
    111  *
    112  * Take a block of UTF-8 chars in and try to convert it to another
    113  * encoding.
    114  * Note: a first call designed to produce heading info is called with
    115  * in = NULL. If stateful this should also initialize the encoder state.
    116  *
    117  * Returns the number of bytes written, -1 if lack of space, or -2
    118  *     if the transcoding failed.
    119  * The value of @inlen after return is the number of octets consumed
    120  *     if the return value is positive, else unpredictiable.
    121  * The value of @outlen after return is the number of octets produced.
    122  */
    123 typedef int (* xmlCharEncodingOutputFunc)(unsigned char *out, int *outlen,
    124                                           const unsigned char *in, int *inlen);
    125 
    126 
    127 /*
    128  * Block defining the handlers for non UTF-8 encodings.
    129  * If iconv is supported, there are two extra fields.
    130  */
    131 #ifdef LIBXML_ICU_ENABLED
    132 struct _uconv_t {
    133   UConverter *uconv; /* for conversion between an encoding and UTF-16 */
    134   UConverter *utf8; /* for conversion between UTF-8 and UTF-16 */
    135 };
    136 typedef struct _uconv_t uconv_t;
    137 #endif
    138 
    139 typedef struct _xmlCharEncodingHandler xmlCharEncodingHandler;
    140 typedef xmlCharEncodingHandler *xmlCharEncodingHandlerPtr;
    141 struct _xmlCharEncodingHandler {
    142     char                       *name;
    143     xmlCharEncodingInputFunc   input;
    144     xmlCharEncodingOutputFunc  output;
    145 #ifdef LIBXML_ICONV_ENABLED
    146     iconv_t                    iconv_in;
    147     iconv_t                    iconv_out;
    148 #endif /* LIBXML_ICONV_ENABLED */
    149 #ifdef LIBXML_ICU_ENABLED
    150     uconv_t                    *uconv_in;
    151     uconv_t                    *uconv_out;
    152 #endif /* LIBXML_ICU_ENABLED */
    153 };
    154 
    155 #ifdef __cplusplus
    156 }
    157 #endif
    158 #include <libxml/tree.h>
    159 #ifdef __cplusplus
    160 extern "C" {
    161 #endif
    162 
    163 /*
    164  * Interfaces for encoding handlers.
    165  */
    166 XMLPUBFUN void XMLCALL
    167 	xmlInitCharEncodingHandlers	(void);
    168 XMLPUBFUN void XMLCALL
    169 	xmlCleanupCharEncodingHandlers	(void);
    170 XMLPUBFUN void XMLCALL
    171 	xmlRegisterCharEncodingHandler	(xmlCharEncodingHandlerPtr handler);
    172 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
    173 	xmlGetCharEncodingHandler	(xmlCharEncoding enc);
    174 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
    175 	xmlFindCharEncodingHandler	(const char *name);
    176 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
    177 	xmlNewCharEncodingHandler	(const char *name,
    178                           		 xmlCharEncodingInputFunc input,
    179                           		 xmlCharEncodingOutputFunc output);
    180 
    181 /*
    182  * Interfaces for encoding names and aliases.
    183  */
    184 XMLPUBFUN int XMLCALL
    185 	xmlAddEncodingAlias		(const char *name,
    186 					 const char *alias);
    187 XMLPUBFUN int XMLCALL
    188 	xmlDelEncodingAlias		(const char *alias);
    189 XMLPUBFUN const char * XMLCALL
    190 	xmlGetEncodingAlias		(const char *alias);
    191 XMLPUBFUN void XMLCALL
    192 	xmlCleanupEncodingAliases	(void);
    193 XMLPUBFUN xmlCharEncoding XMLCALL
    194 	xmlParseCharEncoding		(const char *name);
    195 XMLPUBFUN const char * XMLCALL
    196 	xmlGetCharEncodingName		(xmlCharEncoding enc);
    197 
    198 /*
    199  * Interfaces directly used by the parsers.
    200  */
    201 XMLPUBFUN xmlCharEncoding XMLCALL
    202 	xmlDetectCharEncoding		(const unsigned char *in,
    203 					 int len);
    204 
    205 XMLPUBFUN int XMLCALL
    206 	xmlCharEncOutFunc		(xmlCharEncodingHandler *handler,
    207 					 xmlBufferPtr out,
    208 					 xmlBufferPtr in);
    209 
    210 XMLPUBFUN int XMLCALL
    211 	xmlCharEncInFunc		(xmlCharEncodingHandler *handler,
    212 					 xmlBufferPtr out,
    213 					 xmlBufferPtr in);
    214 XMLPUBFUN int XMLCALL
    215 	xmlCharEncFirstLine		(xmlCharEncodingHandler *handler,
    216 					 xmlBufferPtr out,
    217 					 xmlBufferPtr in);
    218 XMLPUBFUN int XMLCALL
    219 	xmlCharEncCloseFunc		(xmlCharEncodingHandler *handler);
    220 
    221 /*
    222  * Export a few useful functions
    223  */
    224 #ifdef LIBXML_OUTPUT_ENABLED
    225 XMLPUBFUN int XMLCALL
    226 	UTF8Toisolat1			(unsigned char *out,
    227 					 int *outlen,
    228 					 const unsigned char *in,
    229 					 int *inlen);
    230 #endif /* LIBXML_OUTPUT_ENABLED */
    231 XMLPUBFUN int XMLCALL
    232 	isolat1ToUTF8			(unsigned char *out,
    233 					 int *outlen,
    234 					 const unsigned char *in,
    235 					 int *inlen);
    236 #ifdef __cplusplus
    237 }
    238 #endif
    239 
    240 #endif /* __XML_CHAR_ENCODING_H__ */
    241