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. */
29 #include "unicodeio.h"
42 #define _(msgid) gettext (msgid)
43 #define N_(msgid) msgid
45 #include "localcharset.h"
47 /* When we pass a Unicode character to iconv(), we must pass it in a
48 suitable encoding. The standardized Unicode encodings are
49 UTF-8, UCS-2, UCS-4, UTF-16, UTF-16BE, UTF-16LE, UTF-7.
50 UCS-2 supports only characters up to \U0000FFFF.
51 UTF-16 and variants support only characters up to \U0010FFFF.
52 UTF-7 is way too complex and not supported by glibc-2.1.
53 UCS-4 specification leaves doubts about endianness and byte order
54 mark. glibc currently interprets it as big endian without byte order
55 mark, but this is not backed by an RFC.
56 So we use UTF-8. It supports characters up to \U7FFFFFFF and is
57 unambiguously defined. */
59 /* Stores the UTF-8 representation of the Unicode character wc in r[0..5].
60 Returns the number of bytes stored, or -1 if wc is out of range. */
62 utf8_wctomb (unsigned char *r, unsigned int wc)
70 else if (wc < 0x10000)
72 else if (wc < 0x200000)
74 else if (wc < 0x4000000)
76 else if (wc <= 0x7fffffff)
83 /* Note: code falls through cases! */
84 case 6: r[5] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x4000000;
85 case 5: r[4] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x200000;
86 case 4: r[3] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x10000;
87 case 3: r[2] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0x800;
88 case 2: r[1] = 0x80 | (wc & 0x3f); wc = wc >> 6; wc |= 0xc0;
95 /* Luckily, the encoding's name is platform independent. */
96 #define UTF8_NAME "UTF-8"
98 /* Converts the Unicode character CODE to its multibyte representation
99 in the current locale and calls the SUCCESS callback on the resulting
100 byte sequence. If an error occurs, invokes the FAILURE callback instead,
101 passing it CODE and an English error string.
102 Returns whatever the callback returned.
103 Assumes that the locale doesn't change between two calls. */
105 unicode_to_mb (unsigned int code,
106 long (*success) (const char *buf, size_t buflen,
108 long (*failure) (unsigned int code, const char *msg,
112 static int initialized;
115 static iconv_t utf8_to_local;
123 const char *charset = locale_charset ();
125 is_utf8 = !strcmp (charset, UTF8_NAME);
129 utf8_to_local = iconv_open (charset, UTF8_NAME);
130 if (utf8_to_local == (iconv_t)(-1))
131 /* For an unknown encoding, assume ASCII. */
132 utf8_to_local = iconv_open ("ASCII", UTF8_NAME);
138 /* Test whether the utf8_to_local converter is available at all. */
142 if (utf8_to_local == (iconv_t)(-1))
143 return failure (code, N_("iconv function not usable"), callback_arg);
145 return failure (code, N_("iconv function not available"), callback_arg);
149 /* Convert the character to UTF-8. */
150 count = utf8_wctomb ((unsigned char *) inbuf, code);
152 return failure (code, N_("character out of range"), callback_arg);
167 outbytesleft = sizeof (outbuf);
169 /* Convert the character from UTF-8 to the locale's charset. */
170 res = iconv (utf8_to_local,
171 (ICONV_CONST char **)&inptr, &inbytesleft,
172 &outptr, &outbytesleft);
173 if (inbytesleft > 0 || res == (size_t)(-1)
174 /* Irix iconv() inserts a NUL byte if it cannot convert. */
175 # if !defined _LIBICONV_VERSION && (defined sgi || defined __sgi)
176 || (res > 0 && code != 0 && outptr - outbuf == 1 && *outbuf == '\0')
179 return failure (code, NULL, callback_arg);
181 /* Avoid glibc-2.1 bug and Solaris 7 bug. */
182 # if defined _LIBICONV_VERSION \
183 || !((__GLIBC__ - 0 == 2 && __GLIBC_MINOR__ - 0 <= 1) || defined __sun)
185 /* Get back to the initial shift state. */
186 res = iconv (utf8_to_local, NULL, NULL, &outptr, &outbytesleft);
187 if (res == (size_t)(-1))
188 return failure (code, NULL, callback_arg);
191 return success (outbuf, outptr - outbuf, callback_arg);
195 /* At this point, is_utf8 is true, so no conversion is needed. */
196 return success (inbuf, count, callback_arg);
199 /* Simple success callback that outputs the converted string.
200 The STREAM is passed as callback_arg. */
202 fwrite_success_callback (const char *buf, size_t buflen, void *callback_arg)
204 FILE *stream = (FILE *) callback_arg;
206 fwrite (buf, 1, buflen, stream);
210 /* Simple failure callback that displays an error and exits. */
212 exit_failure_callback (unsigned int code, const char *msg, void *callback_arg)
215 error (1, 0, _("cannot convert U+%04X to local character set"), code);
217 error (1, 0, _("cannot convert U+%04X to local character set: %s"), code,
222 /* Simple failure callback that displays a fallback representation in plain
223 ASCII, using the same notation as ISO C99 strings. */
225 fallback_failure_callback (unsigned int code, const char *msg, void *callback_arg)
227 FILE *stream = (FILE *) callback_arg;
230 fprintf (stream, "\\u%04X", code);
232 fprintf (stream, "\\U%08X", code);
236 /* Outputs the Unicode character CODE to the output stream STREAM.
237 Upon failure, exit if exit_on_error is true, otherwise output a fallback
240 print_unicode_char (FILE *stream, unsigned int code, int exit_on_error)
242 unicode_to_mb (code, fwrite_success_callback,
244 ? exit_failure_callback
245 : fallback_failure_callback,