1 /*****************************************************************************
2 * unicode.c: Unicode <-> locale functions
3 *****************************************************************************
4 * Copyright (C) 2005-2006 the VideoLAN team
5 * Copyright © 2005-2006 Rémi Denis-Courmont
8 * Authors: Rémi Denis-Courmont <rem # videolan.org>
10 * This program is free software; you can redistribute it and/or modify
11 * it under the terms of the GNU General Public License as published by
12 * the Free Software Foundation; either version 2 of the License, or
13 * (at your option) any later version.
15 * This program is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with this program; if not, write to the Free Software
22 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
23 *****************************************************************************/
25 /*****************************************************************************
27 *****************************************************************************/
29 #include <vlc_charset.h>
37 #include <sys/types.h>
44 #ifdef HAVE_SYS_STAT_H
45 # include <sys/stat.h>
57 # define lstat( a, b ) stat(a, b)
61 /* Define this if the OS always use UTF-8 internally */
62 # define ASSUME_UTF8 1
66 # if defined (HAVE_ICONV)
67 /* libiconv is more powerful than Win32 API (it has translit) */
69 # elif defined (WIN32) || defined (UNDER_CE)
72 # error No UTF8 charset conversion implemented on this platform!
76 typedef struct locale_data_t
78 #if defined (USE_ICONV)
81 #elif defined (USE_MB2MB)
87 static locale_data_t from_locale, to_locale;
90 void LocaleInit( vlc_object_t *p_this )
95 if( vlc_current_charset( &psz_charset ) )
97 from_locale.hd = to_locale.hd = (vlc_iconv_t)(-1);
101 char psz_buf[strlen( psz_charset ) + sizeof( "//translit" )];
102 const char *psz_conv;
105 * Still allow non-ASCII characters when the locale is not set.
106 * Western Europeans are being favored for historical reasons.
108 if( strcmp( psz_charset, "ASCII" ) )
110 sprintf( psz_buf, "%s//translit", psz_charset );
114 psz_conv = "ISO-8859-1//translit";
116 vlc_mutex_init( p_this, &from_locale.lock );
117 vlc_mutex_init( p_this, &to_locale.lock );
118 from_locale.hd = vlc_iconv_open( "UTF-8", psz_conv );
119 to_locale.hd = vlc_iconv_open( psz_conv, "UTF-8" );
124 assert( (from_locale.hd == (vlc_iconv_t)(-1))
125 == (to_locale.hd == (vlc_iconv_t)(-1)) );
127 #elif defined (USE_MB2MB)
128 to_locale.toCP = from_locale.fromCP = CP_ACP;
129 from_locale.toCP = to_locale.fromCP = CP_UTF8;
135 void LocaleDeinit( void )
138 if( to_locale.hd != (vlc_iconv_t)(-1) )
140 vlc_iconv_close( to_locale.hd );
141 vlc_mutex_destroy( &to_locale.lock );
144 if( from_locale.hd != (vlc_iconv_t)(-1) )
146 vlc_iconv_close( from_locale.hd );
147 vlc_mutex_destroy( &from_locale.lock );
152 static char *locale_fast (const char *string, locale_data_t *p)
154 #if defined (USE_ICONV)
155 vlc_iconv_t hd = p->hd;
157 if (hd == (vlc_iconv_t)(-1))
158 return (char *)string;
160 const char *iptr = string;
161 size_t inb = strlen (string);
162 size_t outb = inb * 6 + 1;
163 char output[outb], *optr = output;
168 vlc_mutex_lock (&p->lock);
169 vlc_iconv (hd, NULL, NULL, NULL, NULL);
171 while (vlc_iconv (hd, &iptr, &inb, &optr, &outb) == (size_t)(-1))
177 vlc_iconv (hd, NULL, NULL, NULL, NULL);
179 vlc_mutex_unlock (&p->lock);
183 assert (*iptr == '\0');
184 assert (*optr == '\0');
185 assert (strlen (output) == (size_t)(optr - output));
186 return strdup (output);
187 #elif defined (USE_MB2MB)
195 len = MultiByteToWideChar (p->fromCP, 0, string, -1, NULL, 0);
201 MultiByteToWideChar (p->fromCP, 0, string, -1, wide, len);
202 len = WideCharToMultiByte (p->toCP, 0, wide, -1, NULL, 0, NULL, NULL);
207 WideCharToMultiByte (p->toCP, 0, wide, -1, out, len, NULL, NULL);
210 return (char *)string;
215 static inline char *locale_dup (const char *string, locale_data_t *p)
217 #if defined (USE_ICONV)
218 return (p->hd == (vlc_iconv_t)(-1))
220 : locale_fast (string, p);
221 #elif defined (USE_MB2MB)
222 return locale_fast (string, p);
224 return strdup (string);
229 void LocaleFree (const char *str)
231 #if defined (USE_ICONV)
232 assert ((to_locale.hd == (vlc_iconv_t)(-1))
233 == (from_locale.hd == (vlc_iconv_t)(-1)));
235 if( to_locale.hd != (vlc_iconv_t)(-1) )
237 #elif defined (USE_MB2MB)
244 * FromLocale: converts a locale string to UTF-8
246 * @param locale nul-terminated string to be converted
248 * @return a nul-terminated UTF-8 string, or NULL in case of error.
249 * To avoid memory leak, you have to pass the result to LocaleFree()
250 * when it is no longer needed.
252 char *FromLocale (const char *locale)
254 return locale_fast (locale, &from_locale);
257 char *FromLocaleDup (const char *locale)
259 return locale_dup (locale, &from_locale);
264 * ToLocale: converts a UTF-8 string to local system encoding.
266 * @param utf8 nul-terminated string to be converted
268 * @return a nul-terminated string, or NULL in case of error.
269 * To avoid memory leak, you have to pass the result to LocaleFree()
270 * when it is no longer needed.
272 char *ToLocale (const char *utf8)
274 return locale_fast (utf8, &to_locale);
278 static char *ToLocaleDup (const char *utf8)
280 return locale_dup (utf8, &to_locale);
285 * utf8_open: open() wrapper for UTF-8 filenames
287 int utf8_open (const char *filename, int flags, mode_t mode)
289 #if defined (WIN32) || defined (UNDER_CE)
290 if (GetVersion() < 0x80000000)
292 /* for Windows NT and above */
293 wchar_t wpath[MAX_PATH + 1];
295 if (!MultiByteToWideChar (CP_UTF8, 0, filename, -1, wpath, MAX_PATH))
300 wpath[MAX_PATH] = L'\0';
303 * open() cannot open files with non-“ANSI” characters on Windows.
304 * We use _wopen() instead. Same thing for mkdir() and stat().
306 return _wopen (wpath, flags, mode);
309 const char *local_name = ToLocale (filename);
311 if (local_name == NULL)
317 int fd = open (local_name, flags, mode);
318 LocaleFree (local_name);
323 * utf8_fopen: fopen() wrapper for UTF-8 filenames
325 FILE *utf8_fopen (const char *filename, const char *mode)
327 int rwflags = 0, oflags = 0;
328 vlc_bool_t append = VLC_FALSE;
330 for (const char *ptr = mode; *ptr; ptr++)
346 oflags |= O_CREAT | O_TRUNC;
361 int fd = utf8_open (filename, rwflags | oflags, 0666);
365 if (append && (lseek (fd, 0, SEEK_END) == -1))
371 FILE *stream = fdopen (fd, mode);
379 * utf8_mkdir: Calls mkdir() after conversion of file name to OS locale
381 * @param dirname a UTF-8 string with the name of the directory that you
383 * @return A 0 return value indicates success. A -1 return value indicates an
384 * error, and an error code is stored in errno
386 int utf8_mkdir( const char *dirname )
388 #if defined (UNDER_CE) || defined (WIN32)
389 wchar_t wname[MAX_PATH + 1];
390 char mod[MAX_PATH + 1];
393 /* Convert '/' into '\' */
394 for( i = 0; *dirname; i++ )
397 return -1; /* overflow */
399 if( *dirname == '/' )
408 if( MultiByteToWideChar( CP_UTF8, 0, mod, -1, wname, MAX_PATH ) == 0 )
413 wname[MAX_PATH] = L'\0';
415 if( CreateDirectoryW( wname, NULL ) == 0 )
417 if( GetLastError( ) == ERROR_ALREADY_EXISTS )
425 char *locname = ToLocale( dirname );
428 if( locname == NULL )
433 res = mkdir( locname, 0755 );
435 LocaleFree( locname );
441 * utf8_opendir: wrapper that converts dirname to the locale in use by the OS
443 * @param dirname UTF-8 representation of the directory name
445 * @return a pointer to the DIR struct. Release with closedir().
447 DIR *utf8_opendir( const char *dirname )
450 wchar_t wname[MAX_PATH + 1];
452 if (MultiByteToWideChar (CP_UTF8, 0, dirname, -1, wname, MAX_PATH))
454 wname[MAX_PATH] = L'\0';
455 return (DIR *)vlc_wopendir (wname);
458 const char *local_name = ToLocale( dirname );
460 if( local_name != NULL )
462 DIR *dir = opendir( local_name );
463 LocaleFree( local_name );
473 * utf8_readdir: a readdir wrapper that returns the name of the next entry
474 * in the directory as a UTF-8 string.
476 * @param dir The directory that is being read
478 * @return a UTF-8 string of the directory entry. Use LocaleFree() to free this memory
480 char *utf8_readdir( DIR *dir )
483 struct _wdirent *ent = vlc_wreaddir (dir);
487 return FromWide (ent->d_name);
491 ent = readdir( (DIR *)dir );
495 return vlc_fix_readdir( ent->d_name );
499 static int dummy_select( const char *str )
505 int utf8_loaddir( DIR *dir, char ***namelist,
506 int (*select)( const char * ),
507 int (*compar)( const char **, const char ** ) )
510 select = dummy_select;
522 while( ( entry = utf8_readdir( dir ) ) != NULL )
526 if( !select( entry ) )
532 newtab = realloc( tab, sizeof( char * ) * (num + 1) );
543 qsort( tab, num, sizeof( tab[0] ),
544 (int (*)( const void *, const void *))compar );
552 for( i = 0; i < num; i++ )
561 int utf8_scandir( const char *dirname, char ***namelist,
562 int (*select)( const char * ),
563 int (*compar)( const char **, const char ** ) )
565 DIR *dir = utf8_opendir (dirname);
570 val = utf8_loaddir (dir, namelist, select, compar);
576 static int utf8_statEx( const char *filename, struct stat *buf,
579 #if defined (WIN32) || defined (UNDER_CE)
580 /* retrieve Windows OS version */
581 if( GetVersion() < 0x80000000 )
583 /* for Windows NT and above */
584 wchar_t wpath[MAX_PATH + 1];
586 if( !MultiByteToWideChar( CP_UTF8, 0, filename, -1, wpath, MAX_PATH ) )
591 wpath[MAX_PATH] = L'\0';
593 return _wstati64( wpath, buf );
596 #ifdef HAVE_SYS_STAT_H
597 const char *local_name = ToLocale( filename );
599 if( local_name != NULL )
601 int res = deref ? stat( local_name, buf )
602 : lstat( local_name, buf );
603 LocaleFree( local_name );
612 int utf8_stat( const char *filename, struct stat *buf)
614 return utf8_statEx( filename, buf, VLC_TRUE );
617 int utf8_lstat( const char *filename, struct stat *buf)
619 return utf8_statEx( filename, buf, VLC_FALSE );
623 * utf8_*printf: *printf with conversion from UTF-8 to local encoding
625 static int utf8_vasprintf( char **str, const char *fmt, va_list ap )
628 int res = vasprintf( &utf8, fmt, ap );
632 *str = ToLocaleDup( utf8 );
637 int utf8_vfprintf( FILE *stream, const char *fmt, va_list ap )
640 int res = utf8_vasprintf( &str, fmt, ap );
644 fputs( str, stream );
649 int utf8_fprintf( FILE *stream, const char *fmt, ... )
655 res = utf8_vfprintf( stream, fmt, ap );
661 static char *CheckUTF8( char *str, char rep )
663 uint8_t *ptr = (uint8_t *)str;
664 assert (str != NULL);
674 for (int i = 0; i < 7; i++)
675 if ((c >> (7 - i)) == ((0xff >> (7 - i)) ^ 1))
683 case 0: // 7-bit ASCII character -> OK
687 case -1: // 1111111x -> error
688 case 1: // continuation byte -> error
692 assert (charlen >= 2);
694 uint32_t cp = c & ~((0xff >> (7 - charlen)) << (7 - charlen));
695 for (int i = 1; i < charlen; i++)
697 assert (cp < (1 << 26));
700 if ((c == '\0') // unexpected end of string
701 || ((c >> 6) != 2)) // not a continuation byte
704 cp = (cp << 6) | (ptr[i] & 0x3f);
707 if (cp < 128) // overlong (special case for ASCII)
709 if (cp < (1u << (5 * charlen - 3))) // overlong
726 * EnsureUTF8: replaces invalid/overlong UTF-8 sequences with question marks
727 * Note that it is not possible to convert from Latin-1 to UTF-8 on the fly,
728 * so we don't try that, even though it would be less disruptive.
730 * @return str if it was valid UTF-8, NULL if not.
732 char *EnsureUTF8( char *str )
734 return CheckUTF8( str, '?' );
739 * IsUTF8: checks whether a string is a valid UTF-8 byte sequence.
741 * @param str nul-terminated string to be checked
743 * @return str if it was valid UTF-8, NULL if not.
745 const char *IsUTF8( const char *str )
747 return CheckUTF8( (char *)str, 0 );