Added support for #pragma code_page.

dlls/commdlg/cdlg_Zh.rc

@@ -17,6 +17,7 @@
  */
 
 LANGUAGE LANG_CHINESE, SUBLANG_CHINESE_TRADITIONAL
+#pragma code_page(936)  /* FIXME: default for CHINESE_TRADITIONAL is 950 */
 
 OPEN_FILE DIALOG LOADONCALL MOVEABLE DISCARDABLE 36, 24, 275, 134
 STYLE DS_MODALFRAME | WS_POPUP | WS_CAPTION | WS_SYSMENU
@@ -220,3 +221,5 @@ STRINGTABLE DISCARDABLE
     IDS_REPORTVIEW       "詳細資料"
     IDS_TODESKTOP        "顯示桌面"
 }
+
+#pragma code_page(default)

dlls/shell32/shell32_Zh.rc

@@ -17,6 +17,7 @@
  */
 
 LANGUAGE LANG_CHINESE, SUBLANG_CHINESE_TRADITIONAL
+#pragma code_page(936)  /* FIXME: default for CHINESE_TRADITIONAL is 950 */
 
 SHELL_ABOUT_MSGBOX DIALOG LOADONCALL MOVEABLE DISCARDABLE 15, 40, 210, 152
 STYLE DS_MODALFRAME | WS_POPUP | WS_VISIBLE | WS_CAPTION | WS_SYSMENU
@@ -55,3 +56,5 @@ BEGIN
 	IDS_SHV_COLUMN6		"使用空間"
 	IDS_SHV_COLUMN7		"剩餘空間"
 END
+
+#pragma code_page(default)

dlls/user/resources/user32_Zh.rc

@@ -76,6 +76,7 @@ STRINGTABLE DISCARDABLE
 }
 
 LANGUAGE LANG_CHINESE, SUBLANG_CHINESE_TRADITIONAL
+#pragma code_page(936)  /* FIXME: default for CHINESE_TRADITIONAL is 950 */
 
 SYSMENU MENU LOADONCALL MOVEABLE DISCARDABLE
 {
@@ -133,3 +134,5 @@ STRINGTABLE DISCARDABLE
 {
     MDI_IDS_MOREWINDOWS  "更多視窗(&M)..." /* defined in mdi.h */
 }
+
+#pragma code_page(default)

tools/wpp/ppy.y

@@ -284,7 +284,7 @@ preprocessor
 	| tGCCLINE tNL		/* The null-token */
 	| tERROR opt_text tNL	{ pperror("#error directive: '%s'", $2); if($2) free($2); }
 	| tWARNING opt_text tNL	{ ppwarning("#warning directive: '%s'", $2); if($2) free($2); }
-	| tPRAGMA opt_text tNL	{ if(pp_status.pedantic) ppwarning("#pragma ignored (arg: '%s')", $2); if($2) free($2); }
+	| tPRAGMA opt_text tNL	{ fprintf(ppout, "#pragma %s\n", $2 ? $2 : ""); if ($2) free($2); }
 	| tPPIDENT opt_text tNL	{ if(pp_status.pedantic) ppwarning("#ident ignored (arg: '%s')", $2); if($2) free($2); }
         | tRCINCLUDE tRCINCLUDEPATH {
                 int nl=strlen($2) +3;

tools/wrc/parser.l

@@ -82,6 +82,9 @@
 %x pp_stripp_final
 /* Set when scanning #line style directives */
 %x pp_line
+/* Set when scanning #pragma */
+%x pp_pragma
+%x pp_code_page
 
 %option stack
 %option never-interactive
@@ -100,6 +103,7 @@ cident	[a-zA-Z_][0-9a-zA-Z_]*
 #include <ctype.h>
 #include <assert.h>
 
+#include "wine/unicode.h"
 #include "wrc.h"
 #include "utils.h"
 #include "parser.h"
@@ -115,7 +119,7 @@ cident	[a-zA-Z_][0-9a-zA-Z_]*
 #define YY_USER_ACTION	char_number+=yyleng; wanted_id = want_id; want_id = 0;
 
 static void addcchar(char c);
-static void addwchar(short s);
+static void addwchar(WCHAR s);
 static string_t *get_buffered_cstring(void);
 static string_t *get_buffered_wstring(void);
 static string_t *make_string(char *s);
@@ -123,13 +127,15 @@ static string_t *make_string(char *s);
 static char *cbuffer;		/* Buffers for string collection */
 static int cbufidx;
 static int cbufalloc = 0;
-static short *wbuffer;
+static WCHAR *wbuffer;
 static int wbufidx;
 static int wbufalloc = 0;
 static int stripslevel = 0;	/* Count {} during pp_strips/pp_stripe mode */
 static int stripplevel = 0;	/* Count () during pp_strips mode */
 static int cjunk_tagline;	/* Where did we start stripping (helps error tracking) */
 
+static int current_codepage = -1;  /* use language default */
+
 /*
  * This one is a bit tricky.
  * We set 'want_id' in the parser to get the first
@@ -316,6 +322,7 @@ static struct keyword *iskeyword(char *kw)
 	 * because we only want to know the linenumber and
 	 * filename.
 	 */
+<INITIAL,pp_strips,pp_stripp>^{ws}*\#{ws}*pragma{ws}+	yy_push_state(pp_pragma);
 <INITIAL,pp_strips,pp_stripp>^{ws}*\#{ws}*	yy_push_state(pp_line);
 <pp_line>[^\n]*	{
 		int lineno;
@@ -337,6 +344,23 @@ static struct keyword *iskeyword(char *kw)
 		input_name = xstrdup(fname);
 	}
 
+<pp_pragma>code_page[^\n]*	yyless(9); yy_pop_state(); yy_push_state(pp_code_page);
+<pp_pragma>[^\n]*		yy_pop_state(); if (pedantic) yywarning("Unrecognized #pragma directive '%s'",yytext);
+
+<pp_code_page>\({ws}*default{ws}*\)[^\n]*	current_codepage = -1; yy_pop_state();
+<pp_code_page>\({ws}*[0-9]+{ws}*\)[^\n]* {
+        char *p = yytext;
+        yy_pop_state();
+        while (*p < '0' || *p > '9') p++;
+        current_codepage = strtol( p, NULL, 10 );
+        if (current_codepage && !cp_get_table( current_codepage ))
+        {
+            yyerror("Codepage %d not supported", current_codepage);
+            current_codepage = 0;
+        }
+    }
+<pp_code_page>[^\n]*	yy_pop_state(); yyerror("Malformed #pragma code_page directive");
+
 	/*
 	 * Strip everything until a ';' taking
 	 * into account braces {} for structures,
@@ -418,16 +442,16 @@ L\"			{
 				return tSTRING;
 			}
 <yylstr>\\[0-7]{1,6}	{ /* octal escape sequence */
-				int result;
-				result = strtol(yytext+1, 0, 8);
+				unsigned int result;
+				result = strtoul(yytext+1, 0, 8);
 				if ( result > 0xffff )
 					yyerror("Character constant out of range");
-				addwchar((short)result);
+				addwchar((WCHAR)result);
 			}
 <yylstr>\\x[0-9a-fA-F]{4} {  /* hex escape sequence */
-				int result;
-				result = strtol(yytext+2, 0, 16);
-				addwchar((short)result);
+				unsigned int result;
+				result = strtoul(yytext+2, 0, 16);
+				addwchar((WCHAR)result);
 			}
 <yylstr>\\x[0-9a-fA-F]{1,3} {  yyerror("Invalid hex escape sequence '%s'", yytext); }
 
@@ -441,8 +465,8 @@ L\"			{
 <yylstr>\\v		addwchar('\v');
 <yylstr>\\(\n|.)	addwchar(yytext[1]);
 <yylstr>\\\r\n		addwchar(yytext[2]);
-<yylstr>\"\"		addcchar('\"');		/* "bla""bla"  -> "bla\"bla" */
-<yylstr>\\\"\"		addcchar('\"');		/* "bla\""bla" -> "bla\"bla" */
+<yylstr>\"\"		addwchar('\"');		/* "bla""bla"  -> "bla\"bla" */
+<yylstr>\\\"\"		addwchar('\"');		/* "bla\""bla" -> "bla\"bla" */
 <yylstr>\"{ws}+\"	;			/* "bla" "bla" -> "blabla" */
 <yylstr>[^\\\n\"]+	{
 				char *yptr = yytext;
@@ -596,7 +620,7 @@ static void addcchar(char c)
 	cbuffer[cbufidx++] = c;
 }
 
-static void addwchar(short s)
+static void addwchar(WCHAR s)
 {
 	if(wbufidx >= wbufalloc)
 	{
@@ -605,27 +629,33 @@ static void addwchar(short s)
 		if(wbufalloc > 65536)
 			yywarning("Reallocating wide string buffer larger than 64kB");
 	}
-
-	/*
-	 * BS 08-Aug-1999 FIXME: The '& 0xff' is probably a bug, but I have
-	 * not experienced it yet and I seem to remember that this was for
-	 * a reason. But, as so many things you tend to forget why.
-	 * I guess that there were problems due to the sign extension of
-	 * shorts WRT chars (e.g. 0x80 becomes 0xff80 instead of 0x0080).
-	 * This should then be fixed in the lexer calling the function.
-	 */
-	wbuffer[wbufidx++] = (short)(s & 0xff);
+	wbuffer[wbufidx++] = s;
 }
 
 static string_t *get_buffered_cstring(void)
 {
-	string_t *str = new_string();
-	str->size = cbufidx;
-	str->type = str_char;
-	str->str.cstr = (char *)xmalloc(cbufidx+1);
-	memcpy(str->str.cstr, cbuffer, cbufidx);
-	str->str.cstr[cbufidx] = '\0';
-	return str;
+    string_t *str = new_string();
+
+    str->size = cbufidx;
+    str->type = str_char;
+    str->str.cstr = (char *)xmalloc(cbufidx+1);
+    memcpy(str->str.cstr, cbuffer, cbufidx);
+    str->str.cstr[cbufidx] = '\0';
+
+    if (!current_codepage || current_codepage == -1 || !win32)  /* store as ANSI string */
+    {
+        if (!current_codepage) yyerror("Codepage set to Unicode only, cannot use ASCII string here");
+        return str;
+    }
+    else  /* convert to Unicode before storing */
+    {
+        string_t *str_w = convert_string( str, str_unicode, current_codepage );
+        if (!check_unicode_conversion( str, str_w, current_codepage ))
+            yyerror("String %s does not convert identically to Unicode and back in codepage %d. "
+                    "Try using a Unicode string instead.", str->str.cstr, current_codepage );
+        free_string( str );
+        return str_w;
+    }
 }
 
 static string_t *get_buffered_wstring(void)
@@ -633,8 +663,8 @@ static string_t *get_buffered_wstring(void)
 	string_t *str = new_string();
 	str->size = wbufidx;
 	str->type = str_unicode;
-	str->str.wstr = (short *)xmalloc(2*(wbufidx+1));
-	memcpy(str->str.wstr, wbuffer, wbufidx);
+	str->str.wstr = xmalloc((wbufidx+1)*sizeof(WCHAR));
+	memcpy(str->str.wstr, wbuffer, wbufidx*sizeof(WCHAR));
 	str->str.wstr[wbufidx] = 0;
 	return str;
 }

tools/wrc/parser.y

@@ -531,6 +531,8 @@ resource
 			yywarning("LANGUAGE not supported in 16-bit mode");
 		if(currentlanguage)
 			free(currentlanguage);
+		if (get_language_codepage($3, $5) == -1)
+			yyerror( "Language %04x is not supported", ($5<<10) + $3);
 		currentlanguage = new_language($3, $5);
 		$$ = NULL;
 		chat("Got LANGUAGE %d,%d (0x%04x)", $3, $5, ($5<<10) + $3);
@@ -1752,7 +1754,10 @@ opt_lvc	: /* Empty */		{ $$ = new_lvc(); }
 	 * The conflict is now moved to the expression handling below.
 	 */
 opt_language
-	: tLANGUAGE expr ',' expr	{ $$ = new_language($2, $4); }
+	: tLANGUAGE expr ',' expr	{ $$ = new_language($2, $4);
+					  if (get_language_codepage($2, $4) == -1)
+						yyerror( "Language %04x is not supported", ($4<<10) + $2);
+					}
 	;
 
 opt_characts

tools/wrc/utils.h

@@ -37,16 +37,16 @@ char *xstrdup(const char *str);
 
 int yyerror(const char *s, ...) __attribute__((format (printf, 1, 2)));
 int yywarning(const char *s, ...) __attribute__((format (printf, 1, 2)));
-void internal_error(const char *file, int line, const char *s, ...) __attribute__((format (printf, 3, 4)));
+void internal_error(const char *file, int line, const char *s, ...) __attribute__((format (printf, 3, 4), noreturn));
 void error(const char *s, ...) __attribute__((format (printf, 1, 2)));
 void warning(const char *s, ...) __attribute__((format (printf, 1, 2)));
 void chat(const char *s, ...) __attribute__((format (printf, 1, 2)));
 
 char *dup_basename(const char *name, const char *ext);
-char *dupwstr2cstr(const short *str);
-short *dupcstr2wstr(const char *str);
 int compare_name_id(name_id_t *n1, name_id_t *n2);
-string_t *convert_string(const string_t *str, enum str_e type);
-void set_language(unsigned short lang, unsigned short sublang);
+string_t *convert_string(const string_t *str, enum str_e type, int codepage);
+void free_string( string_t *str );
+int check_unicode_conversion( const string_t *str_a, const string_t *str_w, int codepage );
+int get_language_codepage( unsigned short lang, unsigned short sublang );
 
 #endif

tools/wrc/wrc.c

@@ -363,6 +363,8 @@ int main(int argc,char *argv[])
 			{
 				int lan;
 				lan = strtol(optarg, NULL, 0);
+				if (get_language_codepage(PRIMARYLANGID(lan), SUBLANGID(lan)) == -1)
+					error("Language %04x is not supported",lan);
 				currentlanguage = new_language(PRIMARYLANGID(lan), SUBLANGID(lan));
 			}
 			break;

tools/wrc/wrctypes.h

@@ -99,7 +99,7 @@ typedef struct string {
 	enum str_e	type;
 	union {
 		char *cstr;
-		short *wstr;
+		WCHAR *wstr;
 	} str;
 } string_t;
 

tools/wrc/writeres.c

@@ -243,26 +243,22 @@ static void write_name_str(FILE *fp, name_id_t *nid)
 	else if(!win32 && nid->name.s_name->type == str_unicode)
 	{
 		name_id_t lnid;
-		string_t str;
 
 		lnid.type = name_str;
-		lnid.name.s_name = &str;
-		str.type = str_char;
-		str.str.cstr = dupwstr2cstr(nid->name.s_name->str.wstr);
+		lnid.name.s_name = convert_string( nid->name.s_name, str_char,
+                                                   get_language_codepage(0,0) );
 		write_name_str(fp, &lnid);
-		free(str.str.cstr);
+		free_string( lnid.name.s_name );
 	}
 	else if(win32 && nid->name.s_name->type == str_char)
 	{
 		name_id_t lnid;
-		string_t str;
 
 		lnid.type = name_str;
-		lnid.name.s_name = &str;
-		str.type = str_unicode;
-		str.str.wstr = dupcstr2wstr(nid->name.s_name->str.cstr);
+		lnid.name.s_name = convert_string( nid->name.s_name, str_unicode,
+                                                   get_language_codepage(0,0) );
 		write_name_str(fp, &lnid);
-		free(str.str.wstr);
+		free_string( lnid.name.s_name );
 	}
 	else  if(win32 && nid->name.s_name->type == str_unicode)
 	{