unicode.c 9.17 KB
Newer Older
1
/*****************************************************************************
2
 * unicode.c: Unicode <-> locale functions
3
 *****************************************************************************
4
 * Copyright (C) 2005-2006 the VideoLAN team
5
 * Copyright © 2005-2008 Rémi Denis-Courmont
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
 *
 * Authors: Rémi Denis-Courmont <rem # videolan.org>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
dionoea's avatar
dionoea committed
21
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
22
23
24
25
26
 *****************************************************************************/

/*****************************************************************************
 * Preamble
 *****************************************************************************/
27
28
29
30
#ifdef HAVE_CONFIG_H
# include "config.h"
#endif

31
#include <vlc_common.h>
zorglub's avatar
zorglub committed
32
#include <vlc_charset.h>
33

34
#include <assert.h>
35

36
#include <stdio.h>
37
#include <stdarg.h>
38
#include <stdlib.h>
39
#include <sys/types.h>
40
41
42
#ifdef UNDER_CE
#  include <tchar.h>
#endif
43

44
45
46
47
48
#ifdef __APPLE__
/* Define this if the OS always use UTF-8 internally */
# define ASSUME_UTF8 1
#endif

Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
49
50
51
52
53
54
55
56
#if defined (ASSUME_UTF8)
/* Cool */
#elif defined (WIN32) || defined (UNDER_CE)
# define USE_MB2MB 1
#elif defined (HAVE_ICONV)
# define USE_ICONV 1
#else
# error No UTF8 charset conversion implemented on this platform!
57
58
#endif

59
#if defined (USE_ICONV)
60
61
# include <langinfo.h>
static char charset[sizeof ("CSISO11SWEDISHFORNAMES")] = "";
62

63
static void find_charset_once (void)
64
{
65
66
67
68
    strlcpy (charset, nl_langinfo (CODESET), sizeof (charset));
    if (!strcasecmp (charset, "ASCII")
     || !strcasecmp (charset, "ANSI_X3.4-1968"))
        strcpy (charset, "UTF-8"); /* superset... */
69
70
}

71
static int find_charset (void)
72
{
73
74
    static pthread_once_t once = PTHREAD_ONCE_INIT;
    pthread_once (&once, find_charset_once);
75
    return !strcasecmp (charset, "UTF-8");
76
}
77
78
#endif

79

80
static char *locale_fast (const char *string, bool from)
81
{
82
83
84
    if( string == NULL )
        return NULL;

85
#if defined (USE_ICONV)
86
87
    if (find_charset ())
        return (char *)string;
88

89
90
    vlc_iconv_t hd = vlc_iconv_open (from ? "UTF-8" : charset,
                                     from ? charset : "UTF-8");
91
    if (hd == (vlc_iconv_t)(-1))
92
        return NULL; /* Uho! */
93
94
95
96
97
98
99
100
101
102
103
104

    const char *iptr = string;
    size_t inb = strlen (string);
    size_t outb = inb * 6 + 1;
    char output[outb], *optr = output;

    while (vlc_iconv (hd, &iptr, &inb, &optr, &outb) == (size_t)(-1))
    {
        *optr++ = '?';
        outb--;
        iptr++;
        inb--;
105
        vlc_iconv (hd, NULL, NULL, NULL, NULL); /* reset */
106
107
    }
    *optr = '\0';
108
    vlc_iconv_close (hd);
109
110
111
112
113
114
115

    assert (inb == 0);
    assert (*iptr == '\0');
    assert (*optr == '\0');
    assert (strlen (output) == (size_t)(optr - output));
    return strdup (output);
#elif defined (USE_MB2MB)
116
    char *out;
117
    int len;
118

119
120
    len = 1 + MultiByteToWideChar (from ? CP_ACP : CP_UTF8,
                                   0, string, -1, NULL, 0);
121
122
123
    wchar_t *wide = malloc (len * sizeof (wchar_t));
    if (wide == NULL)
        return NULL;
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
124

damienf's avatar
damienf committed
125
    MultiByteToWideChar (from ? CP_ACP : CP_UTF8, 0, string, -1, wide, len);
126
127
    len = 1 + WideCharToMultiByte (from ? CP_UTF8 : CP_ACP, 0, wide, -1,
                                   NULL, 0, NULL, NULL);
128
    out = malloc (len);
129
130
131
132
    if (out != NULL)
        WideCharToMultiByte (from ? CP_UTF8 : CP_ACP, 0, wide, -1, out, len,
                             NULL, NULL);
    free (wide);
133
    return out;
134
#else
135
    (void)from;
136
137
138
139
140
    return (char *)string;
#endif
}


141
static inline char *locale_dup (const char *string, bool from)
142
{
Rafaël Carré's avatar
Rafaël Carré committed
143
144
    assert( string );

145
#if defined (USE_ICONV)
146
147
148
    if (find_charset ())
        return strdup (string);
    return locale_fast (string, from);
149
#elif defined (USE_MB2MB)
150
    return locale_fast (string, from);
151
#else
152
    (void)from;
153
154
    return strdup (string);
#endif
155
}
156

Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
157
158
159
160
/**
 * Releases (if needed) a localized or uniformized string.
 * @param str non-NULL return value from FromLocale() or ToLocale().
 */
161
162
163
void LocaleFree (const char *str)
{
#if defined (USE_ICONV)
164
    if (!find_charset ())
165
166
167
        free ((char *)str);
#elif defined (USE_MB2MB)
    free ((char *)str);
168
#else
169
    (void)str;
170
#endif
171
172
}

173

Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
174
/**
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
175
 * Converts a string from the system locale character encoding to UTF-8.
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
176
 *
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
177
 * @param locale nul-terminated string to convert
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
178
179
180
181
182
 *
 * @return a nul-terminated UTF-8 string, or NULL in case of error.
 * To avoid memory leak, you have to pass the result to LocaleFree()
 * when it is no longer needed.
 */
183
char *FromLocale (const char *locale)
184
{
185
    return locale_fast (locale, true);
186
187
}

Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
188
189
190
191
192
193
194
195
196
/**
 * converts a string from the system locale character encoding to utf-8,
 * the result is always allocated on the heap.
 *
 * @param locale nul-terminated string to convert
 *
 * @return a nul-terminated utf-8 string, or null in case of error.
 * The result must be freed using free() - as with the strdup() function.
 */
197
char *FromLocaleDup (const char *locale)
198
{
199
    return locale_dup (locale, true);
200
201
202
}


Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
203
/**
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
204
 * ToLocale: converts an UTF-8 string to local system encoding.
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
205
206
207
208
209
210
211
 *
 * @param utf8 nul-terminated string to be converted
 *
 * @return a nul-terminated string, or NULL in case of error.
 * To avoid memory leak, you have to pass the result to LocaleFree()
 * when it is no longer needed.
 */
212
char *ToLocale (const char *utf8)
213
{
214
    return locale_fast (utf8, false);
215
216
}

217

Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
218
219
220
221
222
223
224
225
226
/**
 * converts a string from UTF-8 to the system locale character encoding,
 * the result is always allocated on the heap.
 *
 * @param utf8 nul-terminated string to convert
 *
 * @return a nul-terminated string, or null in case of error.
 * The result must be freed using free() - as with the strdup() function.
 */
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
227
char *ToLocaleDup (const char *utf8)
228
{
229
    return locale_dup (utf8, false);
230
231
}

232
/**
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
233
234
 * Formats an UTF-8 string as vasprintf(), then print it to stdout, with
 * appropriate conversion to local encoding.
235
 */
236
static int utf8_vasprintf( char **str, const char *fmt, va_list ap )
237
{
238
239
240
241
242
243
244
245
    char *utf8;
    int res = vasprintf( &utf8, fmt, ap );
    if( res == -1 )
        return -1;

    *str = ToLocaleDup( utf8 );
    free( utf8 );
    return res;
246
247
}

Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
248
249
250
251
/**
 * Formats an UTF-8 string as vfprintf(), then print it, with
 * appropriate conversion to local encoding.
 */
252
int utf8_vfprintf( FILE *stream, const char *fmt, va_list ap )
253
{
254
255
256
257
258
259
260
261
    char *str;
    int res = utf8_vasprintf( &str, fmt, ap );
    if( res == -1 )
        return -1;

    fputs( str, stream );
    free( str );
    return res;
262
263
}

Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
264
265
266
267
/**
 * Formats an UTF-8 string as fprintf(), then print it, with
 * appropriate conversion to local encoding.
 */
268
269
int utf8_fprintf( FILE *stream, const char *fmt, ... )
{
270
271
    va_list ap;
    int res;
272

273
274
275
276
    va_start( ap, fmt );
    res = utf8_vfprintf( stream, fmt, ap );
    va_end( ap );
    return res;
277
278
}

279
280

static char *CheckUTF8( char *str, char rep )
281
{
282
    uint8_t *ptr = (uint8_t *)str;
283
284
    assert (str != NULL);

285
    for (;;)
286
    {
287
288
289
290
291
292
293
294
        uint8_t c = ptr[0];
        int charlen = -1;

        if (c == '\0')
            break;

        for (int i = 0; i < 7; i++)
            if ((c >> (7 - i)) == ((0xff >> (7 - i)) ^ 1))
295
            {
296
297
                charlen = i;
                break;
298
            }
299
300

        switch (charlen)
301
        {
302
303
304
305
306
307
            case 0: // 7-bit ASCII character -> OK
                ptr++;
                continue;

            case -1: // 1111111x -> error
            case 1: // continuation byte -> error
308
                goto error;
309
        }
310
311
312
313
314

        assert (charlen >= 2);

        uint32_t cp = c & ~((0xff >> (7 - charlen)) << (7 - charlen));
        for (int i = 1; i < charlen; i++)
315
        {
316
317
318
319
320
            assert (cp < (1 << 26));
            c = ptr[i];

            if ((c == '\0') // unexpected end of string
             || ((c >> 6) != 2)) // not a continuation byte
321
                goto error;
322
323

            cp = (cp << 6) | (ptr[i] & 0x3f);
324
        }
325
326
327
328

        if (cp < 128) // overlong (special case for ASCII)
            goto error;
        if (cp < (1u << (5 * charlen - 3))) // overlong
329
330
            goto error;

331
        ptr += charlen;
332
333
        continue;

334
335
    error:
        if (rep == 0)
336
            return NULL;
337
        *ptr++ = rep;
338
        str = NULL;
339
340
341
342
    }

    return str;
}
343

344
/**
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
345
 * Replaces invalid/overlong UTF-8 sequences with question marks.
346
347
348
349
350
351
352
353
354
355
356
357
 * Note that it is not possible to convert from Latin-1 to UTF-8 on the fly,
 * so we don't try that, even though it would be less disruptive.
 *
 * @return str if it was valid UTF-8, NULL if not.
 */
char *EnsureUTF8( char *str )
{
    return CheckUTF8( str, '?' );
}


/**
Rémi Denis-Courmont's avatar
Rémi Denis-Courmont committed
358
 * Checks whether a string is a valid UTF-8 byte sequence.
359
360
361
362
363
364
365
366
367
 *
 * @param str nul-terminated string to be checked
 *
 * @return str if it was valid UTF-8, NULL if not.
 */
const char *IsUTF8( const char *str )
{
    return CheckUTF8( (char *)str, 0 );
}