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 3 of the License, or
8 (at your option) any later version.
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
16 along with this program. If not, see <http://www.gnu.org/licenses/>. */
18 /* Written by Bruno Haible <haible@clisp.cons.org>. */
20 /* Note: This file requires the locale_charset() function. See in
21 libiconv-1.8/libcharset/INTEGRATE for how to obtain it. */
26 #include "unicodeio.h"
39 #define _(msgid) gettext (msgid)
40 #define N_(msgid) msgid
42 #include "localcharset.h"
44 /* When we pass a Unicode character to iconv(), we must pass it in a
45 suitable encoding. The standardized Unicode encodings are
46 UTF-8, UCS-2, UCS-4, UTF-16, UTF-16BE, UTF-16LE, UTF-7.
47 UCS-2 supports only characters up to \U0000FFFF.
48 UTF-16 and variants support only characters up to \U0010FFFF.
49 UTF-7 is way too complex and not supported by glibc-2.1.
50 UCS-4 specification leaves doubts about endianness and byte order
51 mark. glibc currently interprets it as big endian without byte order
52 mark, but this is not backed by an RFC.
53 So we use UTF-8. It supports characters up to \U7FFFFFFF and is
54 unambiguously defined. */
56 /* Stores the UTF-8 representation of the Unicode character wc in r[0..5].
57 Returns the number of bytes stored, or -1 if wc is out of range. */
59 utf8_wctomb (unsigned char *r, unsigned int wc)
67 else if (wc < 0x10000)
69 else if (wc < 0x200000)
71 else if (wc < 0x4000000)
73 else if (wc <= 0x7fffffff)
80 /* Note: code falls through cases! */
81 case 6: r[5] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x4000000;
82 case 5: r[4] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x200000;
83 case 4: r[3] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x10000;
84 case 3: r[2] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x800;
85 case 2: r[1] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0xc0;
92 /* Luckily, the encoding's name is platform independent. */
93 #define UTF8_NAME "UTF-8"
95 /* Converts the Unicode character CODE to its multibyte representation
96 in the current locale and calls the SUCCESS callback on the resulting
97 byte sequence. If an error occurs, invokes the FAILURE callback instead,
98 passing it CODE and an English error string.
99 Returns whatever the callback returned.
100 Assumes that the locale doesn't change between two calls. */
102 unicode_to_mb (unsigned int code,
103 long (*success) (const char *buf, size_t buflen,
105 long (*failure) (unsigned int code, const char *msg,
109 static int initialized;
112 static iconv_t utf8_to_local;
120 const char *charset = locale_charset ();
122 is_utf8 = !strcmp (charset, UTF8_NAME);
126 utf8_to_local = iconv_open (charset, UTF8_NAME);
127 if (utf8_to_local == (iconv_t)(-1))
128 /* For an unknown encoding, assume ASCII. */
129 utf8_to_local = iconv_open ("ASCII", UTF8_NAME);
135 /* Test whether the utf8_to_local converter is available at all. */
139 if (utf8_to_local == (iconv_t)(-1))
140 return failure (code, N_("iconv function not usable"), callback_arg);
142 return failure (code, N_("iconv function not available"), callback_arg);
146 /* Convert the character to UTF-8. */
147 count = utf8_wctomb ((unsigned char *) inbuf, code);
149 return failure (code, N_("character out of range"), callback_arg);
164 outbytesleft = sizeof (outbuf);
166 /* Convert the character from UTF-8 to the locale's charset. */
167 res = iconv (utf8_to_local,
168 (ICONV_CONST char **)&inptr, &inbytesleft,
169 &outptr, &outbytesleft);
170 if (inbytesleft > 0 || res == (size_t)(-1)
171 /* Irix iconv() inserts a NUL byte if it cannot convert. */
172 # if !defined _LIBICONV_VERSION && (defined sgi || defined __sgi)
173 || (res > 0 && code != 0 && outptr - outbuf == 1 && *outbuf == '\0')
176 return failure (code, NULL, callback_arg);
178 /* Avoid glibc-2.1 bug and Solaris 7 bug. */
179 # if defined _LIBICONV_VERSION \
180 || !((__GLIBC__ - 0 == 2 && __GLIBC_MINOR__ - 0 <= 1) || defined __sun)
182 /* Get back to the initial shift state. */
183 res = iconv (utf8_to_local, NULL, NULL, &outptr, &outbytesleft);
184 if (res == (size_t)(-1))
185 return failure (code, NULL, callback_arg);
188 return success (outbuf, outptr - outbuf, callback_arg);
192 /* At this point, is_utf8 is true, so no conversion is needed. */
193 return success (inbuf, count, callback_arg);
196 /* Simple success callback that outputs the converted string.
197 The STREAM is passed as callback_arg. */
199 fwrite_success_callback (const char *buf, size_t buflen, void *callback_arg)
201 FILE *stream = (FILE *) callback_arg;
203 fwrite (buf, 1, buflen, stream);
207 /* Simple failure callback that displays an error and exits. */
209 exit_failure_callback (unsigned int code, const char *msg, void *callback_arg)
212 error (1, 0, _("cannot convert U+%04X to local character set"), code);
214 error (1, 0, _("cannot convert U+%04X to local character set: %s"), code,
219 /* Simple failure callback that displays a fallback representation in plain
220 ASCII, using the same notation as ISO C99 strings. */
222 fallback_failure_callback (unsigned int code, const char *msg, void *callback_arg)
224 FILE *stream = (FILE *) callback_arg;
227 fprintf (stream, "\\u%04X", code);
229 fprintf (stream, "\\U%08X", code);
233 /* Outputs the Unicode character CODE to the output stream STREAM.
234 Upon failure, exit if exit_on_error is true, otherwise output a fallback
237 print_unicode_char (FILE *stream, unsigned int code, int exit_on_error)
239 unicode_to_mb (code, fwrite_success_callback,
241 ? exit_failure_callback
242 : fallback_failure_callback,