1 /* Unicode character output to streams with locale dependent encoding.
3 Copyright (C) 2000-2003, 2006 Free Software Foundation, Inc.
5 This program is free software; you can redistribute it and/or modify
6 it under the terms of the GNU General Public License as published by
7 the Free Software Foundation; either version 2, or (at your option)
10 This program is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 GNU General Public License for more details.
15 You should have received a copy of the GNU General Public License along
16 with this program; if not, write to the Free Software Foundation,
17 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
19 /* Written by Bruno Haible <haible@clisp.cons.org>. */
21 /* Note: This file requires the locale_charset() function. See in
22 libiconv-1.8/libcharset/INTEGRATE for how to obtain it. */
27 #include "unicodeio.h"
40 #define _(msgid) gettext (msgid)
41 #define N_(msgid) msgid
43 #include "localcharset.h"
45 /* When we pass a Unicode character to iconv(), we must pass it in a
46 suitable encoding. The standardized Unicode encodings are
47 UTF-8, UCS-2, UCS-4, UTF-16, UTF-16BE, UTF-16LE, UTF-7.
48 UCS-2 supports only characters up to \U0000FFFF.
49 UTF-16 and variants support only characters up to \U0010FFFF.
50 UTF-7 is way too complex and not supported by glibc-2.1.
51 UCS-4 specification leaves doubts about endianness and byte order
52 mark. glibc currently interprets it as big endian without byte order
53 mark, but this is not backed by an RFC.
54 So we use UTF-8. It supports characters up to \U7FFFFFFF and is
55 unambiguously defined. */
57 /* Stores the UTF-8 representation of the Unicode character wc in r[0..5].
58 Returns the number of bytes stored, or -1 if wc is out of range. */
60 utf8_wctomb (unsigned char *r, unsigned int wc)
68 else if (wc < 0x10000)
70 else if (wc < 0x200000)
72 else if (wc < 0x4000000)
74 else if (wc <= 0x7fffffff)
81 /* Note: code falls through cases! */
82 case 6: r[5] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x4000000;
83 case 5: r[4] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x200000;
84 case 4: r[3] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x10000;
85 case 3: r[2] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x800;
86 case 2: r[1] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0xc0;
93 /* Luckily, the encoding's name is platform independent. */
94 #define UTF8_NAME "UTF-8"
96 /* Converts the Unicode character CODE to its multibyte representation
97 in the current locale and calls the SUCCESS callback on the resulting
98 byte sequence. If an error occurs, invokes the FAILURE callback instead,
99 passing it CODE and an English error string.
100 Returns whatever the callback returned.
101 Assumes that the locale doesn't change between two calls. */
103 unicode_to_mb (unsigned int code,
104 long (*success) (const char *buf, size_t buflen,
106 long (*failure) (unsigned int code, const char *msg,
110 static int initialized;
113 static iconv_t utf8_to_local;
121 const char *charset = locale_charset ();
123 is_utf8 = !strcmp (charset, UTF8_NAME);
127 utf8_to_local = iconv_open (charset, UTF8_NAME);
128 if (utf8_to_local == (iconv_t)(-1))
129 /* For an unknown encoding, assume ASCII. */
130 utf8_to_local = iconv_open ("ASCII", UTF8_NAME);
136 /* Test whether the utf8_to_local converter is available at all. */
140 if (utf8_to_local == (iconv_t)(-1))
141 return failure (code, N_("iconv function not usable"), callback_arg);
143 return failure (code, N_("iconv function not available"), callback_arg);
147 /* Convert the character to UTF-8. */
148 count = utf8_wctomb ((unsigned char *) inbuf, code);
150 return failure (code, N_("character out of range"), callback_arg);
165 outbytesleft = sizeof (outbuf);
167 /* Convert the character from UTF-8 to the locale's charset. */
168 res = iconv (utf8_to_local,
169 (ICONV_CONST char **)&inptr, &inbytesleft,
170 &outptr, &outbytesleft);
171 if (inbytesleft > 0 || res == (size_t)(-1)
172 /* Irix iconv() inserts a NUL byte if it cannot convert. */
173 # if !defined _LIBICONV_VERSION && (defined sgi || defined __sgi)
174 || (res > 0 && code != 0 && outptr - outbuf == 1 && *outbuf == '\0')
177 return failure (code, NULL, callback_arg);
179 /* Avoid glibc-2.1 bug and Solaris 7 bug. */
180 # if defined _LIBICONV_VERSION \
181 || !((__GLIBC__ - 0 == 2 && __GLIBC_MINOR__ - 0 <= 1) || defined __sun)
183 /* Get back to the initial shift state. */
184 res = iconv (utf8_to_local, NULL, NULL, &outptr, &outbytesleft);
185 if (res == (size_t)(-1))
186 return failure (code, NULL, callback_arg);
189 return success (outbuf, outptr - outbuf, callback_arg);
193 /* At this point, is_utf8 is true, so no conversion is needed. */
194 return success (inbuf, count, callback_arg);
197 /* Simple success callback that outputs the converted string.
198 The STREAM is passed as callback_arg. */
200 fwrite_success_callback (const char *buf, size_t buflen, void *callback_arg)
202 FILE *stream = (FILE *) callback_arg;
204 fwrite (buf, 1, buflen, stream);
208 /* Simple failure callback that displays an error and exits. */
210 exit_failure_callback (unsigned int code, const char *msg, void *callback_arg)
213 error (1, 0, _("cannot convert U+%04X to local character set"), code);
215 error (1, 0, _("cannot convert U+%04X to local character set: %s"), code,
220 /* Simple failure callback that displays a fallback representation in plain
221 ASCII, using the same notation as ISO C99 strings. */
223 fallback_failure_callback (unsigned int code, const char *msg, void *callback_arg)
225 FILE *stream = (FILE *) callback_arg;
228 fprintf (stream, "\\u%04X", code);
230 fprintf (stream, "\\U%08X", code);
234 /* Outputs the Unicode character CODE to the output stream STREAM.
235 Upon failure, exit if exit_on_error is true, otherwise output a fallback
238 print_unicode_char (FILE *stream, unsigned int code, int exit_on_error)
240 unicode_to_mb (code, fwrite_success_callback,
242 ? exit_failure_callback
243 : fallback_failure_callback,