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 *****************************************************************************/
33 #include <vlc_charset.h>
34 #include "libvlc.h" /* utf8_mkdir */
42 #include <sys/types.h>
49 #ifdef HAVE_SYS_STAT_H
50 # include <sys/stat.h>
62 # define lstat( a, b ) stat(a, b)
66 /* Define this if the OS always use UTF-8 internally */
67 # define ASSUME_UTF8 1
70 #if defined (ASSUME_UTF8)
72 #elif defined (WIN32) || defined (UNDER_CE)
74 #elif defined (HAVE_ICONV)
77 # error No UTF8 charset conversion implemented on this platform!
80 #if defined (USE_ICONV)
81 static char charset[sizeof ("CSISO11SWEDISHFORNAMES//translit")] = "";
83 static void find_charset_once (void)
86 if (vlc_current_charset (&psz_charset)
87 || (psz_charset == NULL)
88 || (strcmp (psz_charset, "ASCII") == 0)
89 || ((size_t)snprintf (charset, sizeof (charset), "%s//translit",
90 psz_charset) >= sizeof (charset)))
91 strcpy (charset, "UTF-8");
96 static int find_charset (void)
98 static pthread_once_t once = PTHREAD_ONCE_INIT;
99 pthread_once (&once, find_charset_once);
100 return !strcmp (charset, "UTF-8");
105 static char *locale_fast (const char *string, vlc_bool_t from)
107 #if defined (USE_ICONV)
109 return (char *)string;
111 vlc_iconv_t hd = vlc_iconv_open (from ? "UTF-8" : charset,
112 from ? charset : "UTF-8");
113 if (hd == (vlc_iconv_t)(-1))
114 return strdup (string); /* Uho! */
116 const char *iptr = string;
117 size_t inb = strlen (string);
118 size_t outb = inb * 6 + 1;
119 char output[outb], *optr = output;
124 while (vlc_iconv (hd, &iptr, &inb, &optr, &outb) == (size_t)(-1))
130 vlc_iconv (hd, NULL, NULL, NULL, NULL);
133 vlc_iconv_close (hd);
136 assert (*iptr == '\0');
137 assert (*optr == '\0');
138 assert (strlen (output) == (size_t)(optr - output));
139 return strdup (output);
140 #elif defined (USE_MB2MB)
147 len = 1 + MultiByteToWideChar (from ? CP_ACP : CP_UTF8,
148 0, string, -1, NULL, 0);
151 MultiByteToWideChar (from ? CP_ACP : CP_UTF8, 0, string, -1, wide, len);
152 len = 1 + WideCharToMultiByte (from ? CP_UTF8 : CP_ACP, 0, wide, -1, NULL, 0, NULL, NULL);
157 WideCharToMultiByte (from ? CP_UTF8 : CP_ACP, 0, wide, -1, out, len, NULL, NULL);
161 return (char *)string;
166 static inline char *locale_dup (const char *string, vlc_bool_t from)
168 #if defined (USE_ICONV)
170 return strdup (string);
171 return locale_fast (string, from);
172 #elif defined (USE_MB2MB)
173 return locale_fast (string, from);
176 return strdup (string);
181 void LocaleFree (const char *str)
183 #if defined (USE_ICONV)
184 if (!find_charset ())
186 #elif defined (USE_MB2MB)
195 * FromLocale: converts a locale string to UTF-8
197 * @param locale nul-terminated string to be converted
199 * @return a nul-terminated UTF-8 string, or NULL in case of error.
200 * To avoid memory leak, you have to pass the result to LocaleFree()
201 * when it is no longer needed.
203 char *FromLocale (const char *locale)
205 return locale_fast (locale, VLC_TRUE);
208 char *FromLocaleDup (const char *locale)
210 return locale_dup (locale, VLC_TRUE);
215 * ToLocale: converts a UTF-8 string to local system encoding.
217 * @param utf8 nul-terminated string to be converted
219 * @return a nul-terminated string, or NULL in case of error.
220 * To avoid memory leak, you have to pass the result to LocaleFree()
221 * when it is no longer needed.
223 char *ToLocale (const char *utf8)
225 return locale_fast (utf8, VLC_FALSE);
229 static char *ToLocaleDup (const char *utf8)
231 return locale_dup (utf8, VLC_FALSE);
236 * utf8_open: open() wrapper for UTF-8 filenames
238 int utf8_open (const char *filename, int flags, mode_t mode)
240 #if defined (WIN32) || defined (UNDER_CE)
241 if (GetVersion() < 0x80000000)
243 /* for Windows NT and above */
244 wchar_t wpath[MAX_PATH + 1];
246 if (!MultiByteToWideChar (CP_UTF8, 0, filename, -1, wpath, MAX_PATH))
251 wpath[MAX_PATH] = L'\0';
254 * open() cannot open files with non-“ANSI” characters on Windows.
255 * We use _wopen() instead. Same thing for mkdir() and stat().
257 return _wopen (wpath, flags, mode);
260 const char *local_name = ToLocale (filename);
262 if (local_name == NULL)
268 int fd = open (local_name, flags, mode);
269 LocaleFree (local_name);
274 * utf8_fopen: fopen() wrapper for UTF-8 filenames
276 FILE *utf8_fopen (const char *filename, const char *mode)
278 int rwflags = 0, oflags = 0;
279 vlc_bool_t append = VLC_FALSE;
281 for (const char *ptr = mode; *ptr; ptr++)
297 oflags |= O_CREAT | O_TRUNC;
312 int fd = utf8_open (filename, rwflags | oflags, 0666);
316 if (append && (lseek (fd, 0, SEEK_END) == -1))
322 FILE *stream = fdopen (fd, mode);
330 * utf8_mkdir: Calls mkdir() after conversion of file name to OS locale
332 * @param dirname a UTF-8 string with the name of the directory that you
334 * @return A 0 return value indicates success. A -1 return value indicates an
335 * error, and an error code is stored in errno
337 int utf8_mkdir( const char *dirname, mode_t mode )
339 #if defined (UNDER_CE) || defined (WIN32)
340 wchar_t wname[MAX_PATH + 1];
341 char mod[MAX_PATH + 1];
344 /* Convert '/' into '\' */
345 for( i = 0; *dirname; i++ )
348 return -1; /* overflow */
350 if( *dirname == '/' )
359 if( MultiByteToWideChar( CP_UTF8, 0, mod, -1, wname, MAX_PATH ) == 0 )
364 wname[MAX_PATH] = L'\0';
366 if( CreateDirectoryW( wname, NULL ) == 0 )
368 if( GetLastError( ) == ERROR_ALREADY_EXISTS )
376 char *locname = ToLocale( dirname );
379 if( locname == NULL )
384 res = mkdir( locname, mode );
386 LocaleFree( locname );
392 * utf8_opendir: wrapper that converts dirname to the locale in use by the OS
394 * @param dirname UTF-8 representation of the directory name
396 * @return a pointer to the DIR struct. Release with closedir().
398 DIR *utf8_opendir( const char *dirname )
401 wchar_t wname[MAX_PATH + 1];
403 if (MultiByteToWideChar (CP_UTF8, 0, dirname, -1, wname, MAX_PATH))
405 wname[MAX_PATH] = L'\0';
406 return (DIR *)vlc_wopendir (wname);
409 const char *local_name = ToLocale( dirname );
411 if( local_name != NULL )
413 DIR *dir = opendir( local_name );
414 LocaleFree( local_name );
424 * utf8_readdir: a readdir wrapper that returns the name of the next entry
425 * in the directory as a UTF-8 string.
427 * @param dir The directory that is being read
429 * @return a UTF-8 string of the directory entry. Use free() to free this memory.
431 char *utf8_readdir( DIR *dir )
434 struct _wdirent *ent = vlc_wreaddir (dir);
438 return FromWide (ent->d_name);
442 ent = readdir( (DIR *)dir );
446 return vlc_fix_readdir( ent->d_name );
450 static int dummy_select( const char *str )
456 int utf8_loaddir( DIR *dir, char ***namelist,
457 int (*select)( const char * ),
458 int (*compar)( const char **, const char ** ) )
461 select = dummy_select;
473 while( ( entry = utf8_readdir( dir ) ) != NULL )
477 if( !select( entry ) )
483 newtab = realloc( tab, sizeof( char * ) * (num + 1) );
494 qsort( tab, num, sizeof( tab[0] ),
495 (int (*)( const void *, const void *))compar );
503 for( i = 0; i < num; i++ )
512 int utf8_scandir( const char *dirname, char ***namelist,
513 int (*select)( const char * ),
514 int (*compar)( const char **, const char ** ) )
516 DIR *dir = utf8_opendir (dirname);
521 val = utf8_loaddir (dir, namelist, select, compar);
527 static int utf8_statEx( const char *filename, struct stat *buf,
530 #if defined (WIN32) || defined (UNDER_CE)
531 /* retrieve Windows OS version */
532 if( GetVersion() < 0x80000000 )
534 /* for Windows NT and above */
535 wchar_t wpath[MAX_PATH + 1];
537 if( !MultiByteToWideChar( CP_UTF8, 0, filename, -1, wpath, MAX_PATH ) )
542 wpath[MAX_PATH] = L'\0';
544 return _wstati64( wpath, buf );
547 #ifdef HAVE_SYS_STAT_H
548 const char *local_name = ToLocale( filename );
550 if( local_name != NULL )
552 int res = deref ? stat( local_name, buf )
553 : lstat( local_name, buf );
554 LocaleFree( local_name );
563 int utf8_stat( const char *filename, struct stat *buf)
565 return utf8_statEx( filename, buf, VLC_TRUE );
568 int utf8_lstat( const char *filename, struct stat *buf)
570 return utf8_statEx( filename, buf, VLC_FALSE );
574 * utf8_unlink: Calls unlink() after conversion of file name to OS locale
576 * @param filename a UTF-8 string with the name of the file you want to delete.
577 * @return A 0 return value indicates success. A -1 return value indicates an
578 * error, and an error code is stored in errno
580 int utf8_unlink( const char *filename )
582 #if defined (WIN32) || defined (UNDER_CE)
583 if( GetVersion() < 0x80000000 )
585 /* for Windows NT and above */
586 wchar_t wpath[MAX_PATH + 1];
588 if( !MultiByteToWideChar( CP_UTF8, 0, filename, -1, wpath, MAX_PATH ) )
593 wpath[MAX_PATH] = L'\0';
596 * unlink() cannot open files with non-“ANSI” characters on Windows.
597 * We use _wunlink() instead.
599 return _wunlink( wpath );
602 const char *local_name = ToLocale( filename );
604 if( local_name == NULL )
610 int ret = unlink( local_name );
611 LocaleFree( local_name );
618 * utf8_*printf: *printf with conversion from UTF-8 to local encoding
620 static int utf8_vasprintf( char **str, const char *fmt, va_list ap )
623 int res = vasprintf( &utf8, fmt, ap );
627 *str = ToLocaleDup( utf8 );
632 int utf8_vfprintf( FILE *stream, const char *fmt, va_list ap )
635 int res = utf8_vasprintf( &str, fmt, ap );
639 fputs( str, stream );
644 int utf8_fprintf( FILE *stream, const char *fmt, ... )
650 res = utf8_vfprintf( stream, fmt, ap );
656 static char *CheckUTF8( char *str, char rep )
658 uint8_t *ptr = (uint8_t *)str;
659 assert (str != NULL);
669 for (int i = 0; i < 7; i++)
670 if ((c >> (7 - i)) == ((0xff >> (7 - i)) ^ 1))
678 case 0: // 7-bit ASCII character -> OK
682 case -1: // 1111111x -> error
683 case 1: // continuation byte -> error
687 assert (charlen >= 2);
689 uint32_t cp = c & ~((0xff >> (7 - charlen)) << (7 - charlen));
690 for (int i = 1; i < charlen; i++)
692 assert (cp < (1 << 26));
695 if ((c == '\0') // unexpected end of string
696 || ((c >> 6) != 2)) // not a continuation byte
699 cp = (cp << 6) | (ptr[i] & 0x3f);
702 if (cp < 128) // overlong (special case for ASCII)
704 if (cp < (1u << (5 * charlen - 3))) // overlong
721 * EnsureUTF8: replaces invalid/overlong UTF-8 sequences with question marks
722 * Note that it is not possible to convert from Latin-1 to UTF-8 on the fly,
723 * so we don't try that, even though it would be less disruptive.
725 * @return str if it was valid UTF-8, NULL if not.
727 char *EnsureUTF8( char *str )
729 return CheckUTF8( str, '?' );
734 * IsUTF8: checks whether a string is a valid UTF-8 byte sequence.
736 * @param str nul-terminated string to be checked
738 * @return str if it was valid UTF-8, NULL if not.
740 const char *IsUTF8( const char *str )
742 return CheckUTF8( (char *)str, 0 );