* unicode.c: Unicode <-> locale functions
*****************************************************************************
* Copyright (C) 2005-2006 the VideoLAN team
- * Copyright © 2005-2006 Rémi Denis-Courmont
- * $Id$
+ * Copyright © 2005-2008 Rémi Denis-Courmont
*
* Authors: Rémi Denis-Courmont <rem # videolan.org>
*
/*****************************************************************************
* Preamble
*****************************************************************************/
-#include <vlc/vlc.h>
+#ifdef HAVE_CONFIG_H
+# include "config.h"
+#endif
+
+#include <vlc_common.h>
#include <vlc_charset.h>
-#include "libvlc.h" /* utf8_mkdir */
#include <assert.h>
#include <stdio.h>
#include <stdarg.h>
#include <stdlib.h>
-#include <errno.h>
#include <sys/types.h>
-#ifdef HAVE_DIRENT_H
-# include <dirent.h>
-#endif
#ifdef UNDER_CE
# include <tchar.h>
#endif
-#ifdef HAVE_SYS_STAT_H
-# include <sys/stat.h>
-#endif
-#ifdef HAVE_FCNTL_H
-# include <fcntl.h>
-#endif
-#ifdef WIN32
-# include <io.h>
-#else
-# include <unistd.h>
-#endif
-
-#ifndef HAVE_LSTAT
-# define lstat( a, b ) stat(a, b)
-#endif
#ifdef __APPLE__
/* Define this if the OS always use UTF-8 internally */
#endif
#if defined (USE_ICONV)
-static char charset[sizeof ("CSISO11SWEDISHFORNAMES//translit")] = "";
+# include <langinfo.h>
+static char charset[sizeof ("CSISO11SWEDISHFORNAMES")] = "";
static void find_charset_once (void)
{
- char *psz_charset;
- if (vlc_current_charset (&psz_charset)
- || (psz_charset == NULL)
- || (strcmp (psz_charset, "ASCII") == 0)
- || ((size_t)snprintf (charset, sizeof (charset), "%s//translit",
- psz_charset) >= sizeof (charset)))
- strcpy (charset, "UTF-8");
-
- free (psz_charset);
+ strlcpy (charset, nl_langinfo (CODESET), sizeof (charset));
+ if (!strcasecmp (charset, "ASCII")
+ || !strcasecmp (charset, "ANSI_X3.4-1968"))
+ strcpy (charset, "UTF-8"); /* superset... */
}
static int find_charset (void)
{
static pthread_once_t once = PTHREAD_ONCE_INIT;
pthread_once (&once, find_charset_once);
- return !strcmp (charset, "UTF-8");
+ return !strcasecmp (charset, "UTF-8");
}
#endif
-static char *locale_fast (const char *string, vlc_bool_t from)
+static char *locale_fast (const char *string, bool from)
{
+ if( string == NULL )
+ return NULL;
+
#if defined (USE_ICONV)
if (find_charset ())
return (char *)string;
vlc_iconv_t hd = vlc_iconv_open (from ? "UTF-8" : charset,
from ? charset : "UTF-8");
if (hd == (vlc_iconv_t)(-1))
- return strdup (string); /* Uho! */
+ return NULL; /* Uho! */
const char *iptr = string;
size_t inb = strlen (string);
size_t outb = inb * 6 + 1;
char output[outb], *optr = output;
- if (string == NULL)
- return NULL;
-
while (vlc_iconv (hd, &iptr, &inb, &optr, &outb) == (size_t)(-1))
{
*optr++ = '?';
outb--;
iptr++;
inb--;
- vlc_iconv (hd, NULL, NULL, NULL, NULL);
+ vlc_iconv (hd, NULL, NULL, NULL, NULL); /* reset */
}
*optr = '\0';
vlc_iconv_close (hd);
char *out;
int len;
- if (string == NULL)
- return NULL;
-
len = 1 + MultiByteToWideChar (from ? CP_ACP : CP_UTF8,
0, string, -1, NULL, 0);
wchar_t wide[len];
MultiByteToWideChar (from ? CP_ACP : CP_UTF8, 0, string, -1, wide, len);
- len = 1 + WideCharToMultiByte (from ? CP_UTF8 : CP_ACP, 0, wide, -1, NULL, 0, NULL, NULL);
+ len = 1 + WideCharToMultiByte (from ? CP_UTF8 : CP_ACP, 0, wide, -1,
+ NULL, 0, NULL, NULL);
out = malloc (len);
if (out == NULL)
return NULL;
- WideCharToMultiByte (from ? CP_UTF8 : CP_ACP, 0, wide, -1, out, len, NULL, NULL);
+ WideCharToMultiByte (from ? CP_UTF8 : CP_ACP, 0, wide, -1, out, len,
+ NULL, NULL);
return out;
#else
+ (void)from;
return (char *)string;
#endif
}
-static inline char *locale_dup (const char *string, vlc_bool_t from)
+static inline char *locale_dup (const char *string, bool from)
{
+ assert( string );
+
#if defined (USE_ICONV)
if (find_charset ())
return strdup (string);
#elif defined (USE_MB2MB)
return locale_fast (string, from);
#else
+ (void)from;
return strdup (string);
#endif
}
-
+/**
+ * Releases (if needed) a localized or uniformized string.
+ * @param str non-NULL return value from FromLocale() or ToLocale().
+ */
void LocaleFree (const char *str)
{
#if defined (USE_ICONV)
free ((char *)str);
#elif defined (USE_MB2MB)
free ((char *)str);
+#else
+ (void)str;
#endif
}
/**
- * FromLocale: converts a locale string to UTF-8
+ * Converts a string from the system locale character encoding to UTF-8.
*
- * @param locale nul-terminated string to be converted
+ * @param locale nul-terminated string to convert
*
* @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()
*/
char *FromLocale (const char *locale)
{
- return locale_fast (locale, VLC_TRUE);
+ return locale_fast (locale, true);
}
+/**
+ * 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.
+ */
char *FromLocaleDup (const char *locale)
{
- return locale_dup (locale, VLC_TRUE);
+ return locale_dup (locale, true);
}
/**
- * ToLocale: converts a UTF-8 string to local system encoding.
+ * ToLocale: converts an UTF-8 string to local system encoding.
*
* @param utf8 nul-terminated string to be converted
*
*/
char *ToLocale (const char *utf8)
{
- return locale_fast (utf8, VLC_FALSE);
-}
-
-
-static char *ToLocaleDup (const char *utf8)
-{
- return locale_dup (utf8, VLC_FALSE);
-}
-
-
-/**
- * utf8_open: open() wrapper for UTF-8 filenames
- */
-int utf8_open (const char *filename, int flags, mode_t mode)
-{
-#if defined (WIN32) || defined (UNDER_CE)
- if (GetVersion() < 0x80000000)
- {
- /* for Windows NT and above */
- wchar_t wpath[MAX_PATH + 1];
-
- if (!MultiByteToWideChar (CP_UTF8, 0, filename, -1, wpath, MAX_PATH))
- {
- errno = ENOENT;
- return -1;
- }
- wpath[MAX_PATH] = L'\0';
-
- /*
- * open() cannot open files with non-“ANSI” characters on Windows.
- * We use _wopen() instead. Same thing for mkdir() and stat().
- */
- return _wopen (wpath, flags, mode);
- }
-#endif
- const char *local_name = ToLocale (filename);
-
- if (local_name == NULL)
- {
- errno = ENOENT;
- return -1;
- }
-
- int fd = open (local_name, flags, mode);
- LocaleFree (local_name);
- return fd;
-}
-
-/**
- * utf8_fopen: fopen() wrapper for UTF-8 filenames
- */
-FILE *utf8_fopen (const char *filename, const char *mode)
-{
- int rwflags = 0, oflags = 0;
- vlc_bool_t append = VLC_FALSE;
-
- for (const char *ptr = mode; *ptr; ptr++)
- {
- switch (*ptr)
- {
- case 'r':
- rwflags = O_RDONLY;
- break;
-
- case 'a':
- rwflags = O_WRONLY;
- oflags |= O_CREAT;
- append = VLC_TRUE;
- break;
-
- case 'w':
- rwflags = O_WRONLY;
- oflags |= O_CREAT | O_TRUNC;
- break;
-
- case '+':
- rwflags = O_RDWR;
- break;
-
-#ifdef O_TEXT
- case 't':
- oflags |= O_TEXT;
- break;
-#endif
- }
- }
-
- int fd = utf8_open (filename, rwflags | oflags, 0666);
- if (fd == -1)
- return NULL;
-
- if (append && (lseek (fd, 0, SEEK_END) == -1))
- {
- close (fd);
- return NULL;
- }
-
- FILE *stream = fdopen (fd, mode);
- if (stream == NULL)
- close (fd);
-
- return stream;
-}
-
-/**
- * utf8_mkdir: Calls mkdir() after conversion of file name to OS locale
- *
- * @param dirname a UTF-8 string with the name of the directory that you
- * want to create.
- * @return A 0 return value indicates success. A -1 return value indicates an
- * error, and an error code is stored in errno
- */
-int utf8_mkdir( const char *dirname, mode_t mode )
-{
-#if defined (UNDER_CE) || defined (WIN32)
- wchar_t wname[MAX_PATH + 1];
- char mod[MAX_PATH + 1];
- int i;
-
- /* Convert '/' into '\' */
- for( i = 0; *dirname; i++ )
- {
- if( i == MAX_PATH )
- return -1; /* overflow */
-
- if( *dirname == '/' )
- mod[i] = '\\';
- else
- mod[i] = *dirname;
- dirname++;
-
- }
- mod[i] = 0;
-
- if( MultiByteToWideChar( CP_UTF8, 0, mod, -1, wname, MAX_PATH ) == 0 )
- {
- errno = ENOENT;
- return -1;
- }
- wname[MAX_PATH] = L'\0';
-
- if( CreateDirectoryW( wname, NULL ) == 0 )
- {
- if( GetLastError( ) == ERROR_ALREADY_EXISTS )
- errno = EEXIST;
- else
- errno = ENOENT;
- return -1;
- }
- return 0;
-#else
- char *locname = ToLocale( dirname );
- int res;
-
- if( locname == NULL )
- {
- errno = ENOENT;
- return -1;
- }
- res = mkdir( locname, mode );
-
- LocaleFree( locname );
- return res;
-#endif
+ return locale_fast (utf8, false);
}
-/**
- * utf8_opendir: wrapper that converts dirname to the locale in use by the OS
- *
- * @param dirname UTF-8 representation of the directory name
- *
- * @return a pointer to the DIR struct. Release with closedir().
- */
-DIR *utf8_opendir( const char *dirname )
-{
-#ifdef WIN32
- wchar_t wname[MAX_PATH + 1];
-
- if (MultiByteToWideChar (CP_UTF8, 0, dirname, -1, wname, MAX_PATH))
- {
- wname[MAX_PATH] = L'\0';
- return (DIR *)vlc_wopendir (wname);
- }
-#else
- const char *local_name = ToLocale( dirname );
-
- if( local_name != NULL )
- {
- DIR *dir = opendir( local_name );
- LocaleFree( local_name );
- return dir;
- }
-#endif
-
- errno = ENOENT;
- return NULL;
-}
/**
- * utf8_readdir: a readdir wrapper that returns the name of the next entry
- * in the directory as a UTF-8 string.
- *
- * @param dir The directory that is being read
+ * converts a string from UTF-8 to the system locale character encoding,
+ * the result is always allocated on the heap.
*
- * @return a UTF-8 string of the directory entry. Use free() to free this memory.
- */
-char *utf8_readdir( DIR *dir )
-{
-#ifdef WIN32
- struct _wdirent *ent = vlc_wreaddir (dir);
- if (ent == NULL)
- return NULL;
-
- return FromWide (ent->d_name);
-#else
- struct dirent *ent;
-
- ent = readdir( (DIR *)dir );
- if( ent == NULL )
- return NULL;
-
- return vlc_fix_readdir( ent->d_name );
-#endif
-}
-
-static int dummy_select( const char *str )
-{
- (void)str;
- return 1;
-}
-
-int utf8_loaddir( DIR *dir, char ***namelist,
- int (*select)( const char * ),
- int (*compar)( const char **, const char ** ) )
-{
- if( select == NULL )
- select = dummy_select;
-
- if( dir == NULL )
- return -1;
- else
- {
- char **tab = NULL;
- char *entry;
- unsigned num = 0;
-
- rewinddir( dir );
-
- while( ( entry = utf8_readdir( dir ) ) != NULL )
- {
- char **newtab;
-
- if( !select( entry ) )
- {
- free( entry );
- continue;
- }
-
- newtab = realloc( tab, sizeof( char * ) * (num + 1) );
- if( newtab == NULL )
- {
- free( entry );
- goto error;
- }
- tab = newtab;
- tab[num++] = entry;
- }
-
- if( compar != NULL )
- qsort( tab, num, sizeof( tab[0] ),
- (int (*)( const void *, const void *))compar );
-
- *namelist = tab;
- return num;
-
- error:{
- unsigned i;
-
- for( i = 0; i < num; i++ )
- free( tab[i] );
- if( tab != NULL )
- free( tab );
- }
- }
- return -1;
-}
-
-int utf8_scandir( const char *dirname, char ***namelist,
- int (*select)( const char * ),
- int (*compar)( const char **, const char ** ) )
-{
- DIR *dir = utf8_opendir (dirname);
- int val = -1;
-
- if (dir != NULL)
- {
- val = utf8_loaddir (dir, namelist, select, compar);
- closedir (dir);
- }
- return val;
-}
-
-static int utf8_statEx( const char *filename, struct stat *buf,
- vlc_bool_t deref )
-{
-#if defined (WIN32) || defined (UNDER_CE)
- /* retrieve Windows OS version */
- if( GetVersion() < 0x80000000 )
- {
- /* for Windows NT and above */
- wchar_t wpath[MAX_PATH + 1];
-
- if( !MultiByteToWideChar( CP_UTF8, 0, filename, -1, wpath, MAX_PATH ) )
- {
- errno = ENOENT;
- return -1;
- }
- wpath[MAX_PATH] = L'\0';
-
- return _wstati64( wpath, buf );
- }
-#endif
-#ifdef HAVE_SYS_STAT_H
- const char *local_name = ToLocale( filename );
-
- if( local_name != NULL )
- {
- int res = deref ? stat( local_name, buf )
- : lstat( local_name, buf );
- LocaleFree( local_name );
- return res;
- }
- errno = ENOENT;
-#endif
- return -1;
-}
-
-
-int utf8_stat( const char *filename, struct stat *buf)
-{
- return utf8_statEx( filename, buf, VLC_TRUE );
-}
-
-int utf8_lstat( const char *filename, struct stat *buf)
-{
- return utf8_statEx( filename, buf, VLC_FALSE );
-}
-
-/**
- * utf8_unlink: Calls unlink() after conversion of file name to OS locale
+ * @param utf8 nul-terminated string to convert
*
- * @param filename a UTF-8 string with the name of the file you want to delete.
- * @return A 0 return value indicates success. A -1 return value indicates an
- * error, and an error code is stored in errno
+ * @return a nul-terminated string, or null in case of error.
+ * The result must be freed using free() - as with the strdup() function.
*/
-int utf8_unlink( const char *filename )
+char *ToLocaleDup (const char *utf8)
{
-#if defined (WIN32) || defined (UNDER_CE)
- if( GetVersion() < 0x80000000 )
- {
- /* for Windows NT and above */
- wchar_t wpath[MAX_PATH + 1];
-
- if( !MultiByteToWideChar( CP_UTF8, 0, filename, -1, wpath, MAX_PATH ) )
- {
- errno = ENOENT;
- return -1;
- }
- wpath[MAX_PATH] = L'\0';
-
- /*
- * unlink() cannot open files with non-“ANSI” characters on Windows.
- * We use _wunlink() instead.
- */
- return _wunlink( wpath );
- }
-#endif
- const char *local_name = ToLocale( filename );
-
- if( local_name == NULL )
- {
- errno = ENOENT;
- return -1;
- }
-
- int ret = unlink( local_name );
- LocaleFree( local_name );
- return ret;
+ return locale_dup (utf8, false);
}
-
-
/**
- * utf8_*printf: *printf with conversion from UTF-8 to local encoding
+ * Formats an UTF-8 string as vasprintf(), then print it to stdout, with
+ * appropriate conversion to local encoding.
*/
static int utf8_vasprintf( char **str, const char *fmt, va_list ap )
{
return res;
}
+/**
+ * Formats an UTF-8 string as vfprintf(), then print it, with
+ * appropriate conversion to local encoding.
+ */
int utf8_vfprintf( FILE *stream, const char *fmt, va_list ap )
{
char *str;
return res;
}
+/**
+ * Formats an UTF-8 string as fprintf(), then print it, with
+ * appropriate conversion to local encoding.
+ */
int utf8_fprintf( FILE *stream, const char *fmt, ... )
{
va_list ap;
}
/**
- * EnsureUTF8: replaces invalid/overlong UTF-8 sequences with question marks
+ * Replaces invalid/overlong UTF-8 sequences with question marks.
* 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.
*
/**
- * IsUTF8: checks whether a string is a valid UTF-8 byte sequence.
+ * Checks whether a string is a valid UTF-8 byte sequence.
*
* @param str nul-terminated string to be checked
*