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 /* Size of pivot buffer, same as icu/source/common/ucnv.cpp CHUNK_SIZE */
133 #define ICU_PIVOT_BUF_SIZE 1024
134 struct _uconv_t {
135   UConverter *uconv; /* for conversion between an encoding and UTF-16 */
136   UConverter *utf8; /* for conversion between UTF-8 and UTF-16 */
137   UChar      pivot_buf[ICU_PIVOT_BUF_SIZE];
138   UChar      *pivot_source;
139   UChar      *pivot_target;
140 };
141 typedef struct _uconv_t uconv_t;
142 #endif
143 
144 typedef struct _xmlCharEncodingHandler xmlCharEncodingHandler;
145 typedef xmlCharEncodingHandler *xmlCharEncodingHandlerPtr;
146 struct _xmlCharEncodingHandler {
147     char                       *name;
148     xmlCharEncodingInputFunc   input;
149     xmlCharEncodingOutputFunc  output;
150 #ifdef LIBXML_ICONV_ENABLED
151     iconv_t                    iconv_in;
152     iconv_t                    iconv_out;
153 #endif /* LIBXML_ICONV_ENABLED */
154 #ifdef LIBXML_ICU_ENABLED
155     uconv_t                    *uconv_in;
156     uconv_t                    *uconv_out;
157 #endif /* LIBXML_ICU_ENABLED */
158 };
159 
160 #ifdef __cplusplus
161 }
162 #endif
163 #include <libxml/tree.h>
164 #ifdef __cplusplus
165 extern "C" {
166 #endif
167 
168 /*
169  * Interfaces for encoding handlers.
170  */
171 XMLPUBFUN void XMLCALL
172 	xmlInitCharEncodingHandlers	(void);
173 XMLPUBFUN void XMLCALL
174 	xmlCleanupCharEncodingHandlers	(void);
175 XMLPUBFUN void XMLCALL
176 	xmlRegisterCharEncodingHandler	(xmlCharEncodingHandlerPtr handler);
177 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
178 	xmlGetCharEncodingHandler	(xmlCharEncoding enc);
179 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
180 	xmlFindCharEncodingHandler	(const char *name);
181 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL
182 	xmlNewCharEncodingHandler	(const char *name,
183 					 xmlCharEncodingInputFunc input,
184 					 xmlCharEncodingOutputFunc output);
185 
186 /*
187  * Interfaces for encoding names and aliases.
188  */
189 XMLPUBFUN int XMLCALL
190 	xmlAddEncodingAlias		(const char *name,
191 					 const char *alias);
192 XMLPUBFUN int XMLCALL
193 	xmlDelEncodingAlias		(const char *alias);
194 XMLPUBFUN const char * XMLCALL
195 	xmlGetEncodingAlias		(const char *alias);
196 XMLPUBFUN void XMLCALL
197 	xmlCleanupEncodingAliases	(void);
198 XMLPUBFUN xmlCharEncoding XMLCALL
199 	xmlParseCharEncoding		(const char *name);
200 XMLPUBFUN const char * XMLCALL
201 	xmlGetCharEncodingName		(xmlCharEncoding enc);
202 
203 /*
204  * Interfaces directly used by the parsers.
205  */
206 XMLPUBFUN xmlCharEncoding XMLCALL
207 	xmlDetectCharEncoding		(const unsigned char *in,
208 					 int len);
209 
210 XMLPUBFUN int XMLCALL
211 	xmlCharEncOutFunc		(xmlCharEncodingHandler *handler,
212 					 xmlBufferPtr out,
213 					 xmlBufferPtr in);
214 
215 XMLPUBFUN int XMLCALL
216 	xmlCharEncInFunc		(xmlCharEncodingHandler *handler,
217 					 xmlBufferPtr out,
218 					 xmlBufferPtr in);
219 XMLPUBFUN int XMLCALL
220 	xmlCharEncFirstLine		(xmlCharEncodingHandler *handler,
221 					 xmlBufferPtr out,
222 					 xmlBufferPtr in);
223 XMLPUBFUN int XMLCALL
224 	xmlCharEncCloseFunc		(xmlCharEncodingHandler *handler);
225 
226 /*
227  * Export a few useful functions
228  */
229 #ifdef LIBXML_OUTPUT_ENABLED
230 XMLPUBFUN int XMLCALL
231 	UTF8Toisolat1			(unsigned char *out,
232 					 int *outlen,
233 					 const unsigned char *in,
234 					 int *inlen);
235 #endif /* LIBXML_OUTPUT_ENABLED */
236 XMLPUBFUN int XMLCALL
237 	isolat1ToUTF8			(unsigned char *out,
238 					 int *outlen,
239 					 const unsigned char *in,
240 					 int *inlen);
241 #ifdef __cplusplus
242 }
243 #endif
244 
245 #endif /* __XML_CHAR_ENCODING_H__ */
246