Unicode port: Generate Unicode LangStrings even if source was an ANSI file

git-svn-id: https://svn.code.sf.net/p/nsis/code/NSIS/trunk@6093 212acab6-be3b-0410-9dea-997c60f758d6

INSTALL

@@ -25,6 +25,11 @@ This is a trimmed version of Appendix "Building NSIS" of the documentation.
 
        set ZLIB_W32=C:\Dev\zlib-1.2.4
 
+ - For Unicode version of NSIS
+
+   * type:
+       scons UNICODE=yes
+
  - Installing
 
    * type:

Source/build.cpp

@@ -95,6 +95,7 @@ CEXEBuild::CEXEBuild() :
   linecnt = 0;
   fp = 0;
   curfilename = 0;
+  curfile_unicode = FALSE;
 
   display_info=1;
   display_script=1;

Source/build.h

@@ -107,7 +107,7 @@ class CEXEBuild {
 
     // process a script (you can process as many scripts as you want,
     // it is as if they are concatenated)
-    int process_script(FILE *fp, const TCHAR *curfilename);
+    int process_script(FILE *filepointer, const TCHAR *filename, BOOL unicode);
     int process_oneline(TCHAR *line, const TCHAR *curfilename, int lineptr);
     
     // you only get to call write_output once, so use it wisely.
@@ -125,6 +125,7 @@ class CEXEBuild {
 
     int linecnt;
     const TCHAR *curfilename;
+    BOOL curfile_unicode;
     FILE *fp;
 
     HWND notify_hwnd;
@@ -166,7 +167,7 @@ class CEXEBuild {
     int parseScript();
     int includeScript(TCHAR *f);
     int MacroExists(const TCHAR *macroname);
-    int LoadLicenseFile(TCHAR *file, TCHAR** pdata, LineParser &line);
+    int LoadLicenseFile(TCHAR *file, TCHAR** pdata, LineParser &line, BOOL* unicode);
 #ifdef NSIS_FIX_DEFINES_IN_STRINGS
     void ps_addtoline(const TCHAR *str, GrowBuf &linedata, StringList &hist, bool bIgnoreDefines = false);
 #else
@@ -319,7 +320,7 @@ class CEXEBuild {
      * return aPS_ERROR.  If this function call is overwriting a set user string,
      * this will return a PS_WARNING.
      */
-    int SetLangString(TCHAR *name, LANGID lang, TCHAR *str);
+    int SetLangString(TCHAR *name, LANGID lang, TCHAR *str, BOOL unicode);
 
     /**
      * Sets the user string to the specific NLF_STRINGS id.

Source/lang.cpp

@@ -478,7 +478,7 @@ int CEXEBuild::DefineInnerLangString(int id, int process/*=-1*/) {
 // @return If the language id, the variable name or string is invalid, it will
 // return a PS_ERROR.  If this function call is overwriting a set user string,
 // this will return a PS_WARNING.
-int CEXEBuild::SetLangString(TCHAR *name, LANGID lang, TCHAR *str) {
+int CEXEBuild::SetLangString(TCHAR *name, LANGID lang, TCHAR *str, BOOL unicode) {
   if (!str || !name) return PS_ERROR;
 
   LanguageTable *table = GetLangTable(lang);
@@ -490,7 +490,7 @@ int CEXEBuild::SetLangString(TCHAR *name, LANGID lang, TCHAR *str) {
   if (pos < 0)
     pos = build_langstrings.add(name, &sn);
 
-  if (table->lang_strings->set(sn, str))
+  if (table->lang_strings->set(sn, unicode ? str : CtoTString2(TtoCString(str),table->nlf.m_uCodePage)))
     return PS_WARNING;
 
   return PS_OK;
@@ -912,7 +912,8 @@ TCHAR SkipComments(FILE *f) {
 
 // NSIS Language File parser
 LanguageTable * CEXEBuild::LoadLangFile(TCHAR *filename) {
-  FILE *f = FOPENTEXT(filename, "r");
+  BOOL unicode;
+  FILE *f = FOPENTEXT2(filename, "r", &unicode);
   if (!f) {
     ERROR_MSG(_T("Error: Can't open language file - \"%s\"!\n"),filename);
     return 0;
@@ -1014,6 +1015,15 @@ LanguageTable * CEXEBuild::LoadLangFile(TCHAR *filename) {
       nlf->m_uCodePage = CP_ACP;
   }
 
+#ifdef _UNICODE
+  if (!unicode) // convert font name from ANSI to Unicode now that we know the language codepage
+  {
+    TCHAR* str = nlf->m_szFont;
+    nlf->m_szFont = _tcsdup(CtoTString2(TtoCString(str),table->nlf.m_uCodePage));
+    free(str);
+  }
+#endif
+
   // Get RTL setting
   nlf->m_szStrings[NLF_RTL] = (TCHAR *)malloc(2*sizeof(TCHAR));
   nlf->m_bRTL = false;
@@ -1121,6 +1131,14 @@ LanguageTable * CEXEBuild::LoadLangFile(TCHAR *filename) {
       else *out = *in;
     }
     *out = 0;
+#ifdef _UNICODE
+    if (!unicode)
+    {
+        TCHAR* str = nlf->m_szStrings[i];
+        nlf->m_szStrings[i] = _tcsdup(CtoTString2(TtoCString(str),table->nlf.m_uCodePage));
+        free(str);
+    }
+#endif
   }
   fclose(f);
 

Source/makenssi.cpp

@@ -197,7 +197,8 @@ static tstring get_home()
 
 static int process_config(CEXEBuild& build, tstring& conf)
 {
-  FILE *cfg=FOPENTEXT(conf.c_str(),"rt");
+  BOOL unicode;
+  FILE *cfg=FOPENTEXT2(conf.c_str(),"rt",&unicode);
   if (cfg)
   {
     if (build.display_script) 
@@ -205,7 +206,7 @@ static int process_config(CEXEBuild& build, tstring& conf)
       _ftprintf(g_output,_T("Processing config: \n"));
       fflush(g_output);
     }
-    int ret=build.process_script(cfg,(TCHAR*)conf.c_str());
+    int ret=build.process_script(cfg,(TCHAR*)conf.c_str(),unicode);
     fclose(cfg);
     if (ret != PS_OK && ret != PS_EOF)
     {
@@ -484,6 +485,7 @@ int _tmain(int argc, TCHAR **argv)
 
       {
         TCHAR sfile[1024];
+        BOOL unicode=FALSE;
         if (!_tcscmp(argv[argpos],_T("-")) && !in_files)
         {
           fp=stdin;
@@ -492,11 +494,11 @@ int _tmain(int argc, TCHAR **argv)
         else
         {
           _tcscpy(sfile,argv[argpos]);
-          fp=FOPENTEXT(sfile,"rt");
+          fp=FOPENTEXT2(sfile,"rt",&unicode);
           if (!fp)
           {
             _stprintf(sfile,_T("%s.nsi"),argv[argpos]);
-            fp=FOPENTEXT(sfile,"rt");
+            fp=FOPENTEXT2(sfile,"rt",&unicode);
             if (!fp)
             {
               if (build.display_errors) 
@@ -522,7 +524,7 @@ int _tmain(int argc, TCHAR **argv)
           _ftprintf(g_output,_T("Processing script file: \"%s\"\n"),sfile);
           fflush(g_output);
         }
-        int ret=build.process_script(fp,sfile);
+        int ret=build.process_script(fp,sfile,unicode);
         if (fp != stdin) fclose(fp);
 
         if (ret != PS_EOF && ret != PS_OK)

Source/script.cpp

@@ -202,11 +202,12 @@ void CEXEBuild::del_date_time_predefines()
 }
 #endif
 
-int CEXEBuild::process_script(FILE *filepointer, const TCHAR *filename)
+int CEXEBuild::process_script(FILE *filepointer, const TCHAR *filename, BOOL unicode)
 {
   linecnt = 0;
   fp = filepointer;
   curfilename = filename;
+  curfile_unicode = unicode;
 
   if (has_called_write_output)
   {
@@ -232,6 +233,7 @@ int CEXEBuild::process_script(FILE *filepointer, const TCHAR *filename)
 
   fp = 0;
   curfilename = 0;
+  curfile_unicode = FALSE;
 
   if (m_linebuild.getlen())
   {
@@ -745,7 +747,8 @@ int CEXEBuild::parseScript()
 int CEXEBuild::includeScript(TCHAR *f)
 {
   SCRIPT_MSG(_T("!include: \"%s\"\n"),f);
-  FILE *incfp=FOPENTEXT(f,"rt");
+  BOOL unicode;
+  FILE *incfp=FOPENTEXT2(f,"rt",&unicode);
   if (!incfp)
   {
     ERROR_MSG(_T("!include: could not open file: \"%s\"\n"),f);
@@ -765,7 +768,9 @@ int CEXEBuild::includeScript(TCHAR *f)
   int last_linecnt=linecnt;
   linecnt=0;
   const TCHAR *last_filename=curfilename;
+  BOOL last_unicode=curfile_unicode;
   curfilename=f;
+  curfile_unicode=unicode;
   FILE *last_fp=fp;
   fp=incfp;
 
@@ -787,6 +792,7 @@ int CEXEBuild::includeScript(TCHAR *f)
 
   linecnt=last_linecnt;
   curfilename=last_filename;
+  curfile_unicode=last_unicode;
   fp=last_fp;
 
   build_include_depth--;
@@ -825,9 +831,9 @@ int CEXEBuild::MacroExists(const TCHAR *macroname)
   return 0;
 }
 
-int CEXEBuild::LoadLicenseFile(TCHAR *file, TCHAR** pdata, LineParser &line) // caller must free *pdata, even on error result
+int CEXEBuild::LoadLicenseFile(TCHAR *file, TCHAR** pdata, LineParser &line, BOOL* unicode) // caller must free *pdata, even on error result
 {
-  FILE *fp=FOPENTEXT(file,"rt");
+  FILE *fp=FOPENTEXT2(file,"rt",unicode);
   if (!fp)
   {
     ERROR_MSG(_T("%s: open failed \"%s\"\n"),line.gettoken_str(0),file);
@@ -1593,7 +1599,7 @@ int CEXEBuild::doCommand(int which_token, LineParser &line)
       TCHAR *name = line.gettoken_str(1);
       LANGID lang = line.gettoken_int(2);
       TCHAR *str = line.gettoken_str(3);
-      int ret = SetLangString(name, lang, str);
+      int ret = SetLangString(name, lang, str, curfile_unicode);
       if (ret == PS_WARNING)
         warning_fl(_T("LangString \"%s\" set multiple times for %d, wasting space"), name, lang);
       else if (ret == PS_ERROR) {
@@ -1620,12 +1626,13 @@ int CEXEBuild::doCommand(int which_token, LineParser &line)
 
       TCHAR *data = NULL;
       MANAGE_WITH(data, free);
+      BOOL unicode;
 
-      int ret = LoadLicenseFile(file, &data, line);
+      int ret = LoadLicenseFile(file, &data, line, &unicode);
       if (ret != PS_OK)
           return ret;
 
-      ret = SetLangString(name, lang, data);
+      ret = SetLangString(name, lang, data, unicode);
       if (ret == PS_WARNING)
         warning_fl(_T("LicenseLangString \"%s\" set multiple times for %d, wasting space"), name, lang);
       else if (ret == PS_ERROR)
@@ -1895,7 +1902,8 @@ int CEXEBuild::doCommand(int which_token, LineParser &line)
 
         if (!idx)
         {
-          int ret = LoadLicenseFile(file, &filedata, line);
+          BOOL unicode;
+          int ret = LoadLicenseFile(file, &filedata, line, &unicode);
           if (ret != PS_OK)
               return ret;
           data = filedata;

Source/tstring.cpp

@@ -19,10 +19,11 @@
 #include "util.h"
 #include <vector>
 
-FILE* FileOpenUnicodeText(const TCHAR* file, const TCHAR* mode)
+FILE* FileOpenUnicodeText(const TCHAR* file, const TCHAR* mode, BOOL* unicode)
 {
 	extern FILE *g_output;
 	CValidateUnicode::FILE_TYPE ftype = CValidateUnicode::UTF_8; // default file format is UTF-8
+    if (unicode) *unicode = TRUE;
 
 	// If we are reading an existing file, check to see what type of file it
 	// is first.
@@ -92,7 +93,7 @@ FILE* FileOpenUnicodeText(const TCHAR* file, const TCHAR* mode)
 			break;
 		default:
 			// Looks like fopen() doesn't support other encodings of Unicode.
-			strMode.append(_T(", ccs=UNICODE"));
+            if (unicode) *unicode = FALSE;
 			break;
 	}
 

Source/tstring.h

@@ -29,20 +29,24 @@ typedef std::wstring     tstring;
 typedef std::wofstream   tofstream;
 typedef std::wifstream   tifstream;
 // Use the following macros to open text files.
-FILE* FileOpenUnicodeText(const TCHAR* file, const TCHAR* mode);
+FILE* FileOpenUnicodeText(const TCHAR* file, const TCHAR* mode, BOOL* unicode);
-#define FOPENTEXT(file, mode) FileOpenUnicodeText(file, _T(mode))
+#define FOPENTEXT(file, mode)           FileOpenUnicodeText(file, _T(mode), NULL)
+#define FOPENTEXT2(file, mode, unicode) FileOpenUnicodeText(file, _T(mode), unicode)
 #else
 typedef std::string      tstring;
 typedef std::ofstream    tofstream;
 typedef std::ifstream    tifstream;
 // Use the following macros to open text files.
-#define FOPENTEXT(file, mode) fopen(file, mode)
+#define FOPENTEXT(file, mode)           fopen(file, mode)
+#define FOPENTEXT2(file, mode, unicode) *unicode=FALSE, fopen(file, mode)
 #endif
 
 #ifndef _UNICODE
 #define CtoTString(str) (str)
+#define CtoTString2(str,cp) (str)
 #define TtoCString(str) (str)
 #else
+#define CtoTString2(str,cp) CtoTString(str,cp)
 
 // This is a helpful little function for converting exceptions or
 // other system type things that come back ANSI and must be
@@ -56,6 +60,12 @@ public:
     m_wStr = (wchar_t*) GlobalAlloc(GPTR, len*sizeof(wchar_t));
     MultiByteToWideChar(CP_ACP, 0, str, -1, m_wStr, len);
   }
+  CtoTString(const char* str, UINT cp)
+  {
+    int len = MultiByteToWideChar(cp, 0, str, -1, NULL, 0);
+    m_wStr = (wchar_t*) GlobalAlloc(GPTR, len*sizeof(wchar_t));
+    MultiByteToWideChar(cp, 0, str, -1, m_wStr, len);
+  }
   CtoTString(const std::string& str)
   {
     int len = MultiByteToWideChar(CP_ACP, 0, str.c_str(), str.length()+1, NULL, 0);