• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
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 #else
30 #ifdef LIBXML_ICU_ENABLED
31 #include <unicode/ucnv.h>
32 #if 0
33 /* Forward-declare UConverter here rather than pulling in <unicode/ucnv.h>
34  * to prevent unwanted ICU symbols being exposed to users of libxml2.
35  * One particular case is Qt4 conflicting on UChar32.
36  */
37 #include <stdint.h>
38 struct UConverter;
39 typedef struct UConverter UConverter;
40 #ifdef _MSC_VER
41 typedef wchar_t UChar;
42 #else
43 typedef uint16_t UChar;
44 #endif
45 #endif
46 #endif
47 #endif
48 #ifdef __cplusplus
49 extern "C" {
50 #endif
51 
52 /*
53  * xmlCharEncoding:
54  *
55  * Predefined values for some standard encodings.
56  * Libxml does not do beforehand translation on UTF8 and ISOLatinX.
57  * It also supports ASCII, ISO-8859-1, and UTF16 (LE and BE) by default.
58  *
59  * Anything else would have to be translated to UTF8 before being
60  * given to the parser itself. The BOM for UTF16 and the encoding
61  * declaration are looked at and a converter is looked for at that
62  * point. If not found the parser stops here as asked by the XML REC. A
63  * converter can be registered by the user using xmlRegisterCharEncodingHandler
64  * but the current form doesn't allow stateful transcoding (a serious
65  * problem agreed !). If iconv has been found it will be used
66  * automatically and allow stateful transcoding, the simplest is then
67  * to be sure to enable iconv and to provide iconv libs for the encoding
68  * support needed.
69  *
70  * Note that the generic "UTF-16" is not a predefined value.  Instead, only
71  * the specific UTF-16LE and UTF-16BE are present.
72  */
73 typedef enum {
74     XML_CHAR_ENCODING_ERROR=   -1, /* No char encoding detected */
75     XML_CHAR_ENCODING_NONE=	0, /* No char encoding detected */
76     XML_CHAR_ENCODING_UTF8=	1, /* UTF-8 */
77     XML_CHAR_ENCODING_UTF16LE=	2, /* UTF-16 little endian */
78     XML_CHAR_ENCODING_UTF16BE=	3, /* UTF-16 big endian */
79     XML_CHAR_ENCODING_UCS4LE=	4, /* UCS-4 little endian */
80     XML_CHAR_ENCODING_UCS4BE=	5, /* UCS-4 big endian */
81     XML_CHAR_ENCODING_EBCDIC=	6, /* EBCDIC uh! */
82     XML_CHAR_ENCODING_UCS4_2143=7, /* UCS-4 unusual ordering */
83     XML_CHAR_ENCODING_UCS4_3412=8, /* UCS-4 unusual ordering */
84     XML_CHAR_ENCODING_UCS2=	9, /* UCS-2 */
85     XML_CHAR_ENCODING_8859_1=	10,/* ISO-8859-1 ISO Latin 1 */
86     XML_CHAR_ENCODING_8859_2=	11,/* ISO-8859-2 ISO Latin 2 */
87     XML_CHAR_ENCODING_8859_3=	12,/* ISO-8859-3 */
88     XML_CHAR_ENCODING_8859_4=	13,/* ISO-8859-4 */
89     XML_CHAR_ENCODING_8859_5=	14,/* ISO-8859-5 */
90     XML_CHAR_ENCODING_8859_6=	15,/* ISO-8859-6 */
91     XML_CHAR_ENCODING_8859_7=	16,/* ISO-8859-7 */
92     XML_CHAR_ENCODING_8859_8=	17,/* ISO-8859-8 */
93     XML_CHAR_ENCODING_8859_9=	18,/* ISO-8859-9 */
94     XML_CHAR_ENCODING_2022_JP=  19,/* ISO-2022-JP */
95     XML_CHAR_ENCODING_SHIFT_JIS=20,/* Shift_JIS */
96     XML_CHAR_ENCODING_EUC_JP=   21,/* EUC-JP */
97     XML_CHAR_ENCODING_ASCII=    22 /* pure ASCII */
98 } xmlCharEncoding;
99 
100 /**
101  * xmlCharEncodingInputFunc:
102  * @out:  a pointer to an array of bytes to store the UTF-8 result
103  * @outlen:  the length of @out
104  * @in:  a pointer to an array of chars in the original encoding
105  * @inlen:  the length of @in
106  *
107  * Take a block of chars in the original encoding and try to convert
108  * it to an UTF-8 block of chars out.
109  *
110  * Returns the number of bytes written, -1 if lack of space, or -2
111  *     if the transcoding failed.
112  * The value of @inlen after return is the number of octets consumed
113  *     if the return value is positive, else unpredictiable.
114  * The value of @outlen after return is the number of octets consumed.
115  */
116 typedef int (* xmlCharEncodingInputFunc)(unsigned char *out, int *outlen,
117                                          const unsigned char *in, int *inlen);
118 
119 
120 /**
121  * xmlCharEncodingOutputFunc:
122  * @out:  a pointer to an array of bytes to store the result
123  * @outlen:  the length of @out
124  * @in:  a pointer to an array of UTF-8 chars
125  * @inlen:  the length of @in
126  *
127  * Take a block of UTF-8 chars in and try to convert it to another
128  * encoding.
129  * Note: a first call designed to produce heading info is called with
130  * in = NULL. If stateful this should also initialize the encoder state.
131  *
132  * Returns the number of bytes written, -1 if lack of space, or -2
133  *     if the transcoding failed.
134  * The value of @inlen after return is the number of octets consumed
135  *     if the return value is positive, else unpredictiable.
136  * The value of @outlen after return is the number of octets produced.
137  */
138 typedef int (* xmlCharEncodingOutputFunc)(unsigned char *out, int *outlen,
139                                           const unsigned char *in, int *inlen);
140 
141 
142 /*
143  * Block defining the handlers for non UTF-8 encodings.
144  * If iconv is supported, there are two extra fields.
145  */
146 #ifdef LIBXML_ICU_ENABLED
147 struct _uconv_t {
148   UConverter *uconv; /* for conversion between an encoding and UTF-16 */
149   UConverter *utf8; /* for conversion between UTF-8 and UTF-16 */
150 };
151 typedef struct _uconv_t uconv_t;
152 #endif
153 
154 typedef struct _xmlCharEncodingHandler xmlCharEncodingHandler;
155 typedef xmlCharEncodingHandler *xmlCharEncodingHandlerPtr;
156 struct _xmlCharEncodingHandler {
157     char                       *name;
158     xmlCharEncodingInputFunc   input;
159     xmlCharEncodingOutputFunc  output;
160 #ifdef LIBXML_ICONV_ENABLED
161     iconv_t                    iconv_in;
162     iconv_t                    iconv_out;
163 #endif /* LIBXML_ICONV_ENABLED */
164 #ifdef LIBXML_ICU_ENABLED
165     uconv_t                    *uconv_in;
166     uconv_t                    *uconv_out;
167 #endif /* LIBXML_ICU_ENABLED */
168 };
169 
170 #ifdef __cplusplus
171 }
172 #endif
173 #include <libxml/tree.h>
174 #ifdef __cplusplus
175 extern "C" {
176 #endif
177 
178 /*
179  * Interfaces for encoding handlers.
180  */
181 XMLPUBFUN void XMLCALL
182 	xmlInitCharEncodingHandlers	(void);
183 XMLPUBFUN void XMLCALL
184 	xmlCleanupCharEncodingHandlers	(void);
185 XMLPUBFUN void XMLCALL
186 	xmlRegisterCharEncodingHandler	(xmlCharEncodingHandlerPtr handler);
187 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
188 	xmlGetCharEncodingHandler	(xmlCharEncoding enc);
189 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
190 	xmlFindCharEncodingHandler	(const char *name);
191 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
192 	xmlNewCharEncodingHandler	(const char *name,
193                           		 xmlCharEncodingInputFunc input,
194                           		 xmlCharEncodingOutputFunc output);
195 
196 /*
197  * Interfaces for encoding names and aliases.
198  */
199 XMLPUBFUN int XMLCALL
200 	xmlAddEncodingAlias		(const char *name,
201 					 const char *alias);
202 XMLPUBFUN int XMLCALL
203 	xmlDelEncodingAlias		(const char *alias);
204 XMLPUBFUN const char * XMLCALL
205 	xmlGetEncodingAlias		(const char *alias);
206 XMLPUBFUN void XMLCALL
207 	xmlCleanupEncodingAliases	(void);
208 XMLPUBFUN xmlCharEncoding XMLCALL
209 	xmlParseCharEncoding		(const char *name);
210 XMLPUBFUN const char * XMLCALL
211 	xmlGetCharEncodingName		(xmlCharEncoding enc);
212 
213 /*
214  * Interfaces directly used by the parsers.
215  */
216 XMLPUBFUN xmlCharEncoding XMLCALL
217 	xmlDetectCharEncoding		(const unsigned char *in,
218 					 int len);
219 
220 XMLPUBFUN int XMLCALL
221 	xmlCharEncOutFunc		(xmlCharEncodingHandler *handler,
222 					 xmlBufferPtr out,
223 					 xmlBufferPtr in);
224 
225 XMLPUBFUN int XMLCALL
226 	xmlCharEncInFunc		(xmlCharEncodingHandler *handler,
227 					 xmlBufferPtr out,
228 					 xmlBufferPtr in);
229 XMLPUBFUN int XMLCALL
230 	xmlCharEncFirstLine		(xmlCharEncodingHandler *handler,
231 					 xmlBufferPtr out,
232 					 xmlBufferPtr in);
233 XMLPUBFUN int XMLCALL
234 	xmlCharEncCloseFunc		(xmlCharEncodingHandler *handler);
235 
236 /*
237  * Export a few useful functions
238  */
239 #ifdef LIBXML_OUTPUT_ENABLED
240 XMLPUBFUN int XMLCALL
241 	UTF8Toisolat1			(unsigned char *out,
242 					 int *outlen,
243 					 const unsigned char *in,
244 					 int *inlen);
245 #endif /* LIBXML_OUTPUT_ENABLED */
246 XMLPUBFUN int XMLCALL
247 	isolat1ToUTF8			(unsigned char *out,
248 					 int *outlen,
249 					 const unsigned char *in,
250 					 int *inlen);
251 #ifdef __cplusplus
252 }
253 #endif
254 
255 #endif /* __XML_CHAR_ENCODING_H__ */
256