2 * charset.h - header file for general character set conversion
6 #ifndef charset_charset_h
7 #define charset_charset_h
12 * Enumeration that lists all the multibyte or single-byte
13 * character sets known to this library.
16 CS_NONE
, /* used for reporting errors, etc */
17 CS_ASCII
, /* ordinary US-ASCII is worth having! */
19 CS_ISO8859_1_X11
, /* X font encoding with VT100 glyphs */
97 * This macro is used to initialise a charset_state structure:
99 * charset_state mystate = CHARSET_INIT_STATE;
101 #define CHARSET_INIT_STATE { 0L, 0L } /* a suitable initialiser */
104 * This external variable contains the same data, but is provided
105 * for easy structure-copy assignment:
107 * mystate = charset_init_state;
109 extern const charset_state charset_init_state
;
112 * Routine to convert a MB/SB character set to Unicode.
114 * This routine accepts some number of bytes, updates a state
115 * variable, and outputs some number of Unicode characters. There
116 * are no guarantees. You can't even guarantee that at most one
117 * Unicode character will be output per byte you feed in; for
118 * example, suppose you're reading UTF-8, you've seen E1 80, and
119 * then you suddenly see FE. Now you need to output _two_ error
120 * characters - one for the incomplete sequence E1 80, and one for
121 * the completely invalid UTF-8 byte FE.
123 * Returns the number of wide characters output; will never output
124 * more than the size of the buffer (as specified on input).
125 * Advances the `input' pointer and decrements `inlen', to indicate
126 * how far along the input string it got.
128 * The sequence of `errlen' wide characters pointed to by `errstr'
129 * will be used to indicate a conversion error. If `errstr' is
130 * NULL, `errlen' will be ignored, and the library will choose
131 * something sensible to do on its own. For Unicode, this will be
132 * U+FFFD (REPLACEMENT CHARACTER).
135 int charset_to_unicode(const char **input
, int *inlen
,
136 wchar_t *output
, int outlen
,
137 int charset
, charset_state
*state
,
138 const wchar_t *errstr
, int errlen
);
141 * Routine to convert Unicode to an MB/SB character set.
143 * This routine accepts some number of Unicode characters, updates
144 * a state variable, and outputs some number of bytes.
146 * Returns the number of bytes output; will never output more than
147 * the size of the buffer (as specified on input), and will never
148 * output a partial MB character. Advances the `input' pointer and
149 * decrements `inlen', to indicate how far along the input string
152 * If `error' is non-NULL and a character is found which cannot be
153 * expressed in the output charset, conversion will terminate at
154 * that character (so `input' points to the offending character)
155 * and `*error' will be set to TRUE; if `error' is non-NULL and no
156 * difficult characters are encountered, `*error' will be set to
157 * FALSE. If `error' is NULL, difficult characters will simply be
160 * If `input' is NULL, this routine will output the necessary bytes
161 * to reset the encoding state in any way which might be required
162 * at the end of an output piece of text.
165 int charset_from_unicode(const wchar_t **input
, int *inlen
,
166 char *output
, int outlen
,
167 int charset
, charset_state
*state
, int *error
);
170 * Convert X11 encoding names to and from our charset identifiers.
172 const char *charset_to_xenc(int charset
);
173 int charset_from_xenc(const char *name
);
176 * Convert MIME encoding names to and from our charset identifiers.
178 const char *charset_to_mimeenc(int charset
);
179 int charset_from_mimeenc(const char *name
);
182 * Convert our own encoding names to and from our charset
185 const char *charset_to_localenc(int charset
);
186 int charset_from_localenc(const char *name
);
187 int charset_localenc_nth(int n
);
190 * Convert Mac OS script/region/font to our charset identifiers.
192 int charset_from_macenc(int script
, int region
, int sysvers
,
193 const char *fontname
);
196 * Upgrade a charset identifier to a superset charset which is
197 * often confused with it. For example, people whose MUAs report
198 * their mail as ASCII or ISO8859-1 often in practice turn out to
199 * be using CP1252 quote characters, so when parsing incoming mail
200 * it is prudent to treat ASCII and ISO8859-1 as aliases for CP1252
201 * - and since it's a superset of both, this will cause no
202 * genuinely correct mail to be parsed wrongly.
204 int charset_upgrade(int charset
);
207 * This function returns TRUE if the input charset is a vaguely
208 * sensible superset of ASCII. That is, it returns FALSE for 7-bit
209 * encoding formats such as HZ and UTF-7.
211 int charset_contains_ascii(int charset
);
214 * This function tries to deduce the CS_* identifier of the charset
215 * used in the current C locale. It falls back to CS_ASCII if it
216 * can't figure it out at all, so it will always return a valid
219 * (Note that you should have already called setlocale(LC_CTYPE,
220 * "") to guarantee that this function will do the right thing.)
222 int charset_from_locale(void);
224 #endif /* charset_charset_h */