encoding.h
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 *
16 * See Copyright for the status of this software.
17 *
18 * daniel@veillard.com
19 */
20
21 #ifndef __XML_CHAR_ENCODING_H__
22 #define __XML_CHAR_ENCODING_H__
23
24 #include
25
26 #ifdef LIBXML_ICONV_ENABLED
27 #include
28 #endif
29 #ifdef __cplusplus
30 extern "C" {
31 #endif
32
33 /**
34 * xmlCharEncoding:
35 *
36 * Predefined values for some standard encodings.
37 * Libxml don't do beforehand translation on UTF8, ISOLatinX.
38 * It also support UTF16 (LE and BE) by default.
39 *
40 * Anything else would have to be translated to UTF8 before being
41 * given to the parser itself. The BOM for UTF16 and the encoding
42 * declaration are looked at and a converter is looked for at that
43 * point. If not found the parser stops here as asked by the XML REC
44 * Converter can be registered by the user using xmlRegisterCharEncodingHandler
45 * but the current form doesn't allow stateful transcoding (a serious
46 * problem agreed !). If iconv has been found it will be used
47 * automatically and allow stateful transcoding, the simplest is then
48 * to be sure to enable icon and to provide iconv libs for the encoding
49 * support needed.
50 */
51 typedef enum { <XML_CHAR_ENCODING_ERROR> <XML_CHAR_ENCODING_NONE> <XML_CHAR_ENCODING_UTF8> <XML_CHAR_ENCODING_UTF16LE> <XML_CHAR_ENCODING_UTF16BE> <XML_CHAR_ENCODING_UCS4LE> <XML_CHAR_ENCODING_UCS4BE> <XML_CHAR_ENCODING_EBCDIC> <XML_CHAR_ENCODING_UCS4_2143> <XML_CHAR_ENCODING_UCS4_3412> <XML_CHAR_ENCODING_UCS2> <XML_CHAR_ENCODING_8859_1> <XML_CHAR_ENCODING_8859_2> <XML_CHAR_ENCODING_8859_3> <XML_CHAR_ENCODING_8859_4> <XML_CHAR_ENCODING_8859_5> <XML_CHAR_ENCODING_8859_6> <XML_CHAR_ENCODING_8859_7> <XML_CHAR_ENCODING_8859_8> <XML_CHAR_ENCODING_8859_9> <XML_CHAR_ENCODING_2022_JP> <XML_CHAR_ENCODING_SHIFT_JIS> <XML_CHAR_ENCODING_EUC_JP> <XML_CHAR_ENCODING_ASCII>
52 XML_CHAR_ENCODING_ERROR= -1, /* No char encoding detected */
53 XML_CHAR_ENCODING_NONE= 0, /* No char encoding detected */
54 XML_CHAR_ENCODING_UTF8= 1, /* UTF-8 */
55 XML_CHAR_ENCODING_UTF16LE= 2, /* UTF-16 little endian */
56 XML_CHAR_ENCODING_UTF16BE= 3, /* UTF-16 big endian */
57 XML_CHAR_ENCODING_UCS4LE= 4, /* UCS-4 little endian */
58 XML_CHAR_ENCODING_UCS4BE= 5, /* UCS-4 big endian */
59 XML_CHAR_ENCODING_EBCDIC= 6, /* EBCDIC uh! */
60 XML_CHAR_ENCODING_UCS4_2143=7, /* UCS-4 unusual ordering */
61 XML_CHAR_ENCODING_UCS4_3412=8, /* UCS-4 unusual ordering */
62 XML_CHAR_ENCODING_UCS2= 9, /* UCS-2 */
63 XML_CHAR_ENCODING_8859_1= 10,/* ISO-8859-1 ISO Latin 1 */
64 XML_CHAR_ENCODING_8859_2= 11,/* ISO-8859-2 ISO Latin 2 */
65 XML_CHAR_ENCODING_8859_3= 12,/* ISO-8859-3 */
66 XML_CHAR_ENCODING_8859_4= 13,/* ISO-8859-4 */
67 XML_CHAR_ENCODING_8859_5= 14,/* ISO-8859-5 */
68 XML_CHAR_ENCODING_8859_6= 15,/* ISO-8859-6 */
69 XML_CHAR_ENCODING_8859_7= 16,/* ISO-8859-7 */
70 XML_CHAR_ENCODING_8859_8= 17,/* ISO-8859-8 */
71 XML_CHAR_ENCODING_8859_9= 18,/* ISO-8859-9 */
72 XML_CHAR_ENCODING_2022_JP= 19,/* ISO-2022-JP */
73 XML_CHAR_ENCODING_SHIFT_JIS=20,/* Shift_JIS */
74 XML_CHAR_ENCODING_EUC_JP= 21,/* EUC-JP */
75 XML_CHAR_ENCODING_ASCII= 22 /* pure ASCII */
76 } xmlCharEncoding; <typedef:xmlCharEncoding>
77
78 /**
79 * xmlCharEncodingInputFunc:
80 * @out: a pointer to an array of bytes to store the UTF-8 result
81 * @outlen: the length of @out
82 * @in: a pointer to an array of chars in the original encoding
83 * @inlen: the length of @in
84 *
85 * Take a block of chars in the original encoding and try to convert
86 * it to an UTF-8 block of chars out.
87 *
88 * Returns the number of byte written, or -1 by lack of space, or -2
89 * if the transcoding failed.
90 * The value of @inlen after return is the number of octets consumed
91 * as the return value is positive, else unpredictiable.
92 * The value of @outlen after return is the number of octets consumed.
93 */
94 typedef int (* xmlCharEncodingInputFunc)(unsigned char *out, int *outlen,
95 const unsigned char *in, int *inlen); <typedef:xmlCharEncodingInputFunc>
96
97
98 /**
99 * xmlCharEncodingOutputFunc:
100 * @out: a pointer to an array of bytes to store the result
101 * @outlen: the length of @out
102 * @in: a pointer to an array of UTF-8 chars
103 * @inlen: the length of @in
104 *
105 * Take a block of UTF-8 chars in and try to convert it to an other
106 * encoding.
107 * Note: a first call designed to produce heading info is called with
108 * in = NULL. If stateful this should also initialize the encoder state.
109 *
110 * Returns the number of byte written, or -1 by lack of space, or -2
111 * if the transcoding failed.
112 * The value of @inlen after return is the number of octets consumed
113 * as the return value is positive, else unpredictiable.
114 * The value of @outlen after return is the number of ocetes consumed.
115 */
116 typedef int (* xmlCharEncodingOutputFunc)(unsigned char *out, int *outlen,
117 const unsigned char *in, int *inlen); <typedef:xmlCharEncodingOutputFunc>
118
119
120 /*
121 * Block defining the handlers for non UTF-8 encodings.
122 * If iconv is supported, there is two extra fields.
123 */
124
125 typedef struct _xmlCharEncodingHandler xmlCharEncodingHandler; <_xmlCharEncodingHandler>
126 typedef xmlCharEncodingHandler *xmlCharEncodingHandlerPtr; <typedef:xmlCharEncodingHandlerPtr>
127 struct _xmlCharEncodingHandler {
128 char *name;
129 xmlCharEncodingInputFunc input;
130 xmlCharEncodingOutputFunc output;
131 #ifdef LIBXML_ICONV_ENABLED
132 iconv_t iconv_in;
133 iconv_t iconv_out;
134 #endif /* LIBXML_ICONV_ENABLED */
135 };
136
137 #ifdef __cplusplus
138 }
139 #endif
140 #include
141 #ifdef __cplusplus
142 extern "C" {
143 #endif
144
145 /*
146 * Interfaces for encoding handlers.
147 */
148 void xmlInitCharEncodingHandlers (void);
149 void xmlCleanupCharEncodingHandlers (void);
150 void xmlRegisterCharEncodingHandler (xmlCharEncodingHandlerPtr handler);
151 xmlCharEncodingHandlerPtr
152 xmlGetCharEncodingHandler (xmlCharEncoding enc);
153 xmlCharEncodingHandlerPtr
154 xmlFindCharEncodingHandler (const char *name);
155 xmlCharEncodingHandlerPtr
156 xmlNewCharEncodingHandler (const char *name,
157 xmlCharEncodingInputFunc input,
158 xmlCharEncodingOutputFunc output);
159
160 /*
161 * Interfaces for encoding names and aliases.
162 */
163 int xmlAddEncodingAlias (const char *name,
164 const char *alias);
165 int xmlDelEncodingAlias (const char *alias);
166 const char *
167 xmlGetEncodingAlias (const char *alias);
168 void xmlCleanupEncodingAliases (void);
169 xmlCharEncoding
170 xmlParseCharEncoding (const char *name);
171 const char *
172 xmlGetCharEncodingName (xmlCharEncoding enc);
173
174 /*
175 * Interfaces directly used by the parsers.
176 */
177 xmlCharEncoding
178 xmlDetectCharEncoding (const unsigned char *in,
179 int len);
180
181 int xmlCharEncOutFunc (xmlCharEncodingHandler *handler,
182 xmlBufferPtr out,
183 xmlBufferPtr in);
184
185 int xmlCharEncInFunc (xmlCharEncodingHandler *handler,
186 xmlBufferPtr out,
187 xmlBufferPtr in);
188 int xmlCharEncFirstLine (xmlCharEncodingHandler *handler,
189 xmlBufferPtr out,
190 xmlBufferPtr in);
191 int xmlCharEncCloseFunc (xmlCharEncodingHandler *handler);
192
193 /*
194 * Export a few useful functions
195 */
196 int UTF8Toisolat1 (unsigned char *out,
197 int *outlen,
198 const unsigned char *in,
199 int *inlen);
200 int isolat1ToUTF8 (unsigned char *out,
201 int *outlen,
202 const unsigned char *in,
203 int *inlen);
204 int xmlGetUTF8Char (const unsigned char *utf,
205 int *len);
206 /*
207 * exports additional "UTF-8 aware" string routines which are.
208 */
209
210 int xmlCheckUTF8 (const unsigned char *utf);
211
212 int xmlUTF8Strsize (const xmlChar *utf,
213 int len);
214 xmlChar * xmlUTF8Strndup (const xmlChar *utf,
215 int len);
216 xmlChar * xmlUTF8Strpos (const xmlChar *utf,
217 int pos);
218 int xmlUTF8Strloc (const xmlChar *utf,
219 const xmlChar *utfchar);
220 xmlChar * xmlUTF8Strsub (const xmlChar *utf,
221 int start,
222 int len);
223
224 int xmlUTF8Strlen (const xmlChar *utf);
225
226 #ifdef __cplusplus
227 }
228 #endif
229
230 #endif /* __XML_CHAR_ENCODING_H__ */