Improve locale implementation (should now behave identically to bash) and document locale variables behaviour

darcs-hash:20060108230049-ac50b-403f1d00c8483fc4fecc275b62e40b1c3d51bfc1.gz

common.c

@@ -104,8 +104,17 @@ static struct winsize termsize;
 */
 static int block_count=0;
 
+static string_buffer_t *setlocale_buff=0;
+
+
 void common_destroy()
 {
+	if( setlocale_buff )
+	{
+		sb_destroy( setlocale_buff );
+		free( setlocale_buff );
+	}
+	
 	debug( 3, L"Calls: wcsdupcat %d, wcsdupcat2 %d, wcsndup %d, str2wcs %d, wcs2str %d", c1, c2, c3, c4, c5 );
 }
 
@@ -719,27 +728,33 @@ wchar_t *quote_end( const wchar_t *in )
 }
 
 
-void fish_setlocale(int category, const wchar_t *locale)
+const wchar_t *wsetlocale(int category, const wchar_t *locale)
 {
-	char *lang = wcs2str( locale );
-	setlocale(category,lang);
+
+	char *lang = locale?wcs2str( locale ):0;
+	char * res = setlocale(category,lang);
 	
 	free( lang );
+
 	/*
 	  Use ellipsis if on known unicode system, otherwise use $
 	*/
-	if( wcslen( locale ) )
+	char *ctype = setlocale( LC_CTYPE, (void *)0 );
+	ellipsis_char = (strstr( ctype, ".UTF")||strstr( ctype, ".utf") )?L'\u2026':L'$';	
+		
+	if( !res )
+		return 0;
+	
+	if( !setlocale_buff )
 	{
-		ellipsis_char = wcsstr( locale, L".UTF")?L'\u2026':L'$';	
-	}
-	else
-	{
-		char *lang = getenv( "LANG" );
-		if( lang )
-			ellipsis_char = strstr( lang, ".UTF")?L'\u2026':L'$';	
-		else
-			ellipsis_char = L'$';
+		setlocale_buff = malloc( sizeof(string_buffer_t) );
+		sb_init( setlocale_buff);
 	}
+	
+	sb_clear( setlocale_buff );
+	sb_printf( setlocale_buff, L"%s", res );
+	
+	return (wchar_t *)setlocale_buff->buff;	
 }
 
 int contains_str( const wchar_t *a, ... )

common.h

@@ -256,9 +256,12 @@ wchar_t *quote_end( const wchar_t *in );
 void error_reset();
 
 /**
-   Set the locale, also change the ellipsis character
+   This function behaves exactly like a wide character equivalent of
+   the C function setlocale, except that it will also try to detect if
+   the user is using a Unicode character set, and if so, use the
+   unicode ellipsis character as ellipsis, instead of '$'.      
 */
-void fish_setlocale( int category, const wchar_t *locale );
+const wchar_t *wsetlocale( int category, const wchar_t *locale );
 
 /**
    Checks if \c needle is included in the list of strings specified

doc_src/doc.hdr

@@ -655,6 +655,7 @@ values of these variables. They are:
 - \c PWD, which is the current working directory. 
 - \c status, which is the exit status of the last foreground job to exit. If a job contains pipelines, the status of the last command in the pipeline is the status for the job.
 - \c USER, which is the username. This variable can only be changed by the root user.
+- \c LANG, \cLC_ALL, \c LC_COLLATE, \c LC_CTYPE, \cLC_MESSAGES, \c LC_MONETARY, \c LC_NUMERIC and LC_TIME set the language option for the shell and subprograms. See the section <a href='#variables-locale'>Locale variables</a> for more information.
 
 Variables whose name are in uppercase are exported to the commands
 started by fish. This rule is not enforced by fish, but it is good
@@ -663,6 +664,22 @@ unexported variables. \c fish also uses several variables
 internally. Such variables are prefixed with the string __FISH or
 __fish. These should be ignored by the user.
 
+\subsection variables-locale Locale variables
+
+The most common way to set the locale to use a command like 'set -x
+LANG en_GB.utf8', which sets the current locale to be the english
+language, adapted to great britain, using the UTF-8 character set. For
+a list of available locales, use 'locale -a'.
+
+\c LANG, \cLC_ALL, \c LC_COLLATE, \c LC_CTYPE, \cLC_MESSAGES, \c
+LC_MONETARY, \c LC_NUMERIC and LC_TIME set the language option for the
+shell and subprograms. These variables work as follows: \c LC_ALL
+forces all the aspects of the locale to the specified value. If LC_ALL
+is set, all other locale variables will be ignored. The other LC_
+variables set the specified aspect of the locale information. . LANG
+is a fallback value, it will be used if none of the LC_ variables are
+specified.
+
 \section builtin-overview Builtins 
 
 Many other shells have a large library of builtin commands. Most of

env.c

@@ -214,6 +214,73 @@ static mode_t get_umask()
 	return res;
 }
 
+/**
+   Checks if the specified variable is a locale variable
+*/
+static int is_locale( const wchar_t *key )
+{
+	return contains_str( key, L"LANG", L"LC_ALL", L"LC_COLLATE", L"LC_CTYPE", L"LC_MESSAGES", L"LC_MONETARY", L"LC_NUMERIC", L"LC_TIME", (void *)0);
+}
+
+/**
+  Properly sets all locale information
+*/
+static void handle_locale()
+{
+	const wchar_t *lc_all = env_get( L"LC_ALL" );
+	const wchar_t *lang;
+	int i;
+	wchar_t *old = wcsdup(wsetlocale( LC_MESSAGES, (void *)0 ));
+	
+	static const wchar_t *lc[] = 
+		{
+			L"LC_COLLATE", L"LC_CTYPE", L"LC_MESSAGES", L"LC_MONETARY", L"LC_NUMERIC", L"LC_TIME", (void *)0
+		}
+	;
+	static const int cat[] = 
+		{
+			LC_COLLATE, LC_CTYPE, LC_MESSAGES, LC_MONETARY, LC_NUMERIC, LC_TIME
+		}
+	;
+	
+	if( lc_all )
+	{
+		wsetlocale( LC_ALL, lc_all );
+	}
+	else
+	{
+		lang = env_get( L"LANG" );
+		if( lang )
+		{
+			wsetlocale( LC_ALL, lang );
+		}
+		
+		for( i=0; lc[i]; i++ )
+		{
+			const wchar_t *val = env_get( lc[i] );
+			if( val )
+				wsetlocale(  cat[i], val );
+		}
+	}
+	
+	if( wcscmp( wsetlocale( LC_MESSAGES, (void *)0 ), old ) != 0 )
+	{
+		/* Make change known to gettext.  */
+		{
+			extern int  _nl_msg_cat_cntr;
+			++_nl_msg_cat_cntr;
+		}			
+		
+		if( is_interactive )
+		{
+			debug( 0, _(L"Changing language to english") );
+		}
+	}
+	free( old );
+		
+}
+
+
 /**
    Universal variable callback function. This function makes sure the
    proper events are triggered when an event occurs.
@@ -224,6 +291,9 @@ static void universal_callback( int type,
 {
 	wchar_t *str=0;
 	
+	if( is_locale( name ) )
+		handle_locale();
+	
 	switch( type )
 	{
 		case SET:
@@ -475,7 +545,7 @@ static env_node_t *env_get_node( const wchar_t *key )
 	
 	return 0;
 }
-	
+
 void env_set( const wchar_t *key, 
 			  const wchar_t *val, 
 			  int var_mode )
@@ -490,26 +560,13 @@ void env_set( const wchar_t *key,
 
 	event_t ev;
 	int is_universal = 0;	
-	
+
 	if( (var_mode & ENV_USER ) && 
 		hash_get( &env_read_only, key ) )
 	{
 		return;
 	}
 	
-	if( wcscmp(key, L"LANG" )==0 )
-	{
-		fish_setlocale(LC_ALL,val);
-		/* Make change known to gettext.  */
-		{
-			extern int  _nl_msg_cat_cntr;
-			++_nl_msg_cat_cntr;
-		}
-
-		if( is_interactive )
-			debug( 0, _(L"Changing language to english") );
-	}
-
 	if( wcscmp( key, L"umask" ) == 0)
 	{
 		wchar_t *end;
@@ -533,7 +590,6 @@ void env_set( const wchar_t *key,
 		*/
 		return;
 	}
-	
 
 	/*
 	  Zero element arrays are internaly not coded as null but as this placeholder string
@@ -685,9 +741,17 @@ void env_set( const wchar_t *key,
 //	debug( 1, L"env_set: return from event firing" );	
 		al_destroy( &ev.arguments );	
 	}
-	
+
+	if( is_locale( key ) )
+	{
+		handle_locale();
+	}
+		
 }
 
+
+
+
 /**
    Attempt to remove/free the specified key/value pair from the
    specified hash table.
@@ -735,6 +799,10 @@ void env_remove( const wchar_t *key, int var_mode )
 	{
 		env_universal_remove( key );
 	}
+
+	if( is_locale( key ) )
+		handle_locale();
+			
 }
 
 

fish_pager.c

@@ -845,8 +845,7 @@ static void init()
 {
 	struct sigaction act;
 	program_name = L"fish_pager";
-	fish_setlocale( LC_ALL, L"" );
-	
+	wsetlocale( LC_ALL, L"" );
 	
 	int out = dup( 1 );
 	close(1);

fishd.c

@@ -400,7 +400,7 @@ static void init()
 
 	sock = get_socket();
 	daemonize();	
-	fish_setlocale( LC_ALL, L"" );	
+	wsetlocale( LC_ALL, L"" );	
 	env_universal_common_init( &broadcast );
 	
 	load();	

parser.c

@@ -2114,7 +2114,7 @@ static void eval_job( tokenizer *tok )
 					   tok_get_desc( tok_last_type(tok) ) );
 			}
 			
-			return 0;
+			return;
 		}
 				
 		case TOK_ERROR: