added wxLocale::GetSystemEncoding

git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@10394 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775

docs/latex/wx/font.tex

@@ -13,6 +13,68 @@ a window's text.
 
 <wx/font.h>
 
+\wxheading{Constants}
+
+\begin{verbatim}
+enum wxFontEncoding
+{
+    wxFONTENCODING_SYSTEM = -1,     // system default
+    wxFONTENCODING_DEFAULT,         // current default encoding
+
+    // ISO8859 standard defines a number of single-byte charsets
+    wxFONTENCODING_ISO8859_1,       // West European (Latin1)
+    wxFONTENCODING_ISO8859_2,       // Central and East European (Latin2)
+    wxFONTENCODING_ISO8859_3,       // Esperanto (Latin3)
+    wxFONTENCODING_ISO8859_4,       // Baltic (old) (Latin4)
+    wxFONTENCODING_ISO8859_5,       // Cyrillic
+    wxFONTENCODING_ISO8859_6,       // Arabic
+    wxFONTENCODING_ISO8859_7,       // Greek
+    wxFONTENCODING_ISO8859_8,       // Hebrew
+    wxFONTENCODING_ISO8859_9,       // Turkish (Latin5)
+    wxFONTENCODING_ISO8859_10,      // Variation of Latin4 (Latin6)
+    wxFONTENCODING_ISO8859_11,      // Thai
+    wxFONTENCODING_ISO8859_12,      // doesn't exist currently, but put it
+                                    // here anyhow to make all ISO8859
+                                    // consecutive numbers
+    wxFONTENCODING_ISO8859_13,      // Baltic (Latin7)
+    wxFONTENCODING_ISO8859_14,      // Latin8
+    wxFONTENCODING_ISO8859_15,      // Latin9 (a.k.a. Latin0, includes euro)
+    wxFONTENCODING_ISO8859_MAX,
+
+    // Cyrillic charset soup (see http://czyborra.com/charsets/cyrillic.html)
+    wxFONTENCODING_KOI8,            // we don't support any of KOI8 variants
+    wxFONTENCODING_ALTERNATIVE,     // same as MS-DOS CP866
+    wxFONTENCODING_BULGARIAN,       // used under Linux in Bulgaria
+
+    // what would we do without Microsoft? They have their own encodings
+        // for DOS
+    wxFONTENCODING_CP437,           // original MS-DOS codepage
+    wxFONTENCODING_CP850,           // CP437 merged with Latin1
+    wxFONTENCODING_CP852,           // CP437 merged with Latin2
+    wxFONTENCODING_CP855,           // another cyrillic encoding
+    wxFONTENCODING_CP866,           // and another one
+        // and for Windows
+    wxFONTENCODING_CP874,           // WinThai
+    wxFONTENCODING_CP1250,          // WinLatin2
+    wxFONTENCODING_CP1251,          // WinCyrillic
+    wxFONTENCODING_CP1252,          // WinLatin1
+    wxFONTENCODING_CP1253,          // WinGreek (8859-7)
+    wxFONTENCODING_CP1254,          // WinTurkish
+    wxFONTENCODING_CP1255,          // WinHebrew
+    wxFONTENCODING_CP1256,          // WinArabic
+    wxFONTENCODING_CP1257,          // WinBaltic (same as Latin 7)
+    wxFONTENCODING_CP12_MAX,
+
+    wxFONTENCODING_UTF7,            // UTF-7 Unicode encoding
+    wxFONTENCODING_UTF8,            // UTF-8 Unicode encoding
+
+    wxFONTENCODING_UNICODE,         // Unicode - currently used only by
+                                    // wxEncodingConverter class
+
+    wxFONTENCODING_MAX
+};
+\end{verbatim}
+
 \wxheading{Predefined objects}
 
 Objects:

docs/latex/wx/fontmap.tex

@@ -31,10 +31,9 @@ for "equivalent" encodings (e.g. iso8859-2 and cp1250) and tries them.
 
 If you need to display text in encoding which is not available at
 host system (see \helpref{IsEncodingAvailable}{wxfontmapperisencodingavailable}),
-you may use these two classes to a) find font in some similar encoding
+you may use these two classes to find font in some similar encoding
 (see \helpref{GetAltForEncoding}{wxfontmappergetaltforencoding})
-and
+and convert the text to this encoding 
-b) convert the text to this encoding 
 (\helpref{wxEncodingConverter::Convert}{wxencodingconverterconvert}).
 
 Following code snippet demonstrates it:

docs/latex/wx/locale.tex

@@ -438,6 +438,26 @@ Returns current platform-specific locale name as passed to setlocale().
 Compare \helpref{GetCanonicalName}{wxlocalegetcanonicalname}.
 
 
+\membersection{wxLocale::GetSystemEncoding}\label{wxlocalegetsystemencoding}
+
+\constfunc{static wxFontEncoding}{GetSystemEncoding}{\void}
+
+Tries to detect the user's default font encoding.
+Returns \helpref{wxFontEncoding}{wxfont} value or 
+{\bf wxFONTENCODING\_SYSTEM} if it couldn't be determined.
+
+\membersection{wxLocale::GetSystemEncodingName}\label{wxlocalegetsystemencodingname}
+
+\constfunc{static wxString}{GetSystemEncodingName}{\void}
+
+Tries to detect the name of the user's default font encoding. This string isn't
+particularly useful for the application as its form is platform-dependent and
+so you should probably use 
+\helpref{GetSystemEncoding}{wxlocalegetsystemencoding} instead.
+
+Returns a user-readable string value or an empty string if it couldn't be
+determined.
+
 \membersection{wxLocale::GetSystemLanguage}\label{wxlocalegetsystemlanguage}
 
 \constfunc{static int}{GetSystemLanguage}{\void}

include/wx/intl.h

@@ -9,8 +9,8 @@
 // Licence:     wxWindows license
 /////////////////////////////////////////////////////////////////////////////
 
-#ifndef   __INTLH__
+#ifndef _WX_INTL_H_
-#define   __INTLH__
+#define _WX_INTL_H_
 
 #ifdef __GNUG__
     #pragma interface "intl.h"
@@ -21,6 +21,8 @@
 
 #if wxUSE_INTL
 
+#include "wx/fontenc.h"
+
 // ============================================================================
 // global decls
 // ============================================================================
@@ -301,8 +303,6 @@ enum wxLanguage
 
 // --- --- --- generated code ends here --- --- ---
 
-
-
 // ----------------------------------------------------------------------------
 // wxLanguageInfo: encapsulates wxLanguage to OS native lang.desc.
 //                 translation information
@@ -313,8 +313,9 @@ struct WXDLLEXPORT wxLanguageInfo
     int Language;                   // wxLanguage id
     wxString CanonicalName;         // Canonical name, e.g. fr_FR
 #ifdef __WIN32__
-    wxUint32 WinLang, WinSublang;   // Win32 language identifiers
+    wxUint32 WinLang,               // Win32 language identifiers
-#endif
+             WinSublang;
+#endif // __WIN32__
     wxString Description;           // human-readable name of the language
 };
 
@@ -329,7 +330,6 @@ enum wxLocaleInitFlags
     wxLOCALE_CONV_ENCODING = 0x0002      // convert encoding on the fly?
 };
 
-
 class WXDLLEXPORT wxLocale
 {
 public:
@@ -370,6 +370,14 @@ public:
     // Return wxLANGUAGE_UNKNOWN if language-guessing algorithm failed
     static int GetSystemLanguage();
 
+    // get the encoding used by default for text on this system, returns
+    // wxFONTENCODING_SYSTEM if it couldn't be determined
+    static wxFontEncoding GetSystemEncoding();
+
+    // get the string describing the system encoding, return empty string if
+    // couldn't be determined
+    static wxString GetSystemEncodingName();
+
     // return TRUE if the locale was set successfully
     bool IsOk() const { return m_pszOldLocale != NULL; }
 
@@ -492,5 +500,5 @@ inline const wxChar *wxGetTranslation(const wxChar *sz) { return sz; }
     #define gettext_noop(str) _T(str)
 #endif
 
-#endif
+#endif // _WX_INTL_H_
-    // _WX_INTLH__
+

include/wx/strconv.h

@@ -23,6 +23,7 @@
 #if defined(__VISAGECPP__) && __IBMCPP__ >= 400
 #  undef __BSEXCPT__
 #endif
+
 #include <stdlib.h>
 
 #if wxUSE_WCHAR_T
@@ -133,6 +134,8 @@ public:
 private:
     void SetName(const wxChar *charset);
 
+    // note that we can't use wxString here because of compilation
+    // dependencies: we're included from wx/string.h
     wxChar *m_name;
     wxCharacterSet *m_cset;
     bool m_deferred;

samples/console/console.cpp

@@ -39,7 +39,7 @@
 //#define TEST_CHARSET
 //#define TEST_CMDLINE
 //#define TEST_DATETIME
-#define TEST_DIR
+//#define TEST_DIR
 //#define TEST_DLLLOADER
 //#define TEST_ENVIRON
 //#define TEST_EXECUTE
@@ -50,7 +50,7 @@
 //#define TEST_HASH
 //#define TEST_INFO_FUNCTIONS
 //#define TEST_LIST
-//#define TEST_LOCALE
+#define TEST_LOCALE
 //#define TEST_LOG
 //#define TEST_LONGLONG
 //#define TEST_MIME
@@ -1173,11 +1173,19 @@ static void TestDefaultLang()
         _T("klingonese"),   // I bet on some systems it does exist...
     };
 
+    wxPrintf(_T("The default system encoding is %s (%d)\n"),
+             wxLocale::GetSystemEncodingName().c_str(),
+             wxLocale::GetSystemEncoding());
+
     for ( size_t n = 0; n < WXSIZEOF(langStrings); n++ )
     {
         const char *langStr = langStrings[n];
         if ( langStr )
+        {
+            // FIXME: this doesn't do anything at all under Windows, we need
+            //        to create a new wxLocale!
             wxSetEnv(_T("LC_ALL"), langStr);
+        }
 
         int lang = gs_localeDefault.GetSystemLanguage();
         printf("Locale for '%s' is %s.\n",

src/common/intl.cpp

@@ -1,5 +1,5 @@
 /////////////////////////////////////////////////////////////////////////////
-// Name:        intl.cpp
+// Name:        src/common/intl.cpp
 // Purpose:     Internationalization and localisation for wxWindows
 // Author:      Vadim Zeitlin
 // Modified by:
@@ -33,25 +33,28 @@
 // standard headers
 #include <locale.h>
 #include <ctype.h>
+#include <stdlib.h>
 
 // wxWindows
-#include "wx/defs.h"
+#ifndef WX_PRECOMP
-#include "wx/string.h"
+    #include "wx/string.h"
-#include "wx/tokenzr.h"
+    #include "wx/intl.h"
-#include "wx/intl.h"
+    #include "wx/log.h"
+    #include "wx/debug.h"
+    #include "wx/utils.h"
+    #include "wx/dynarray.h"
+#endif // WX_PRECOMP
+
 #include "wx/file.h"
-#include "wx/log.h"
+#include "wx/tokenzr.h"
-#include "wx/debug.h"
-#include "wx/utils.h"
-#include "wx/dynarray.h"
 #include "wx/module.h"
+
 #ifdef __WIN32__
-#include "wx/msw/private.h"
+    #include "wx/msw/private.h"
+#elif defined(__UNIX_LIKE__)
+    #include "wx/fontmap.h"         // for CharsetToEncoding()
 #endif
 
-
-#include <stdlib.h>
-
 // ----------------------------------------------------------------------------
 // simple types
 // ----------------------------------------------------------------------------
@@ -930,6 +933,87 @@ void wxLocale::AddCatalogLookupPathPrefix(const wxString& prefix)
     return wxLANGUAGE_UNKNOWN;
 }
 
+// ----------------------------------------------------------------------------
+// encoding stuff
+// ----------------------------------------------------------------------------
+
+// this is a bit strange as under Windows we get the encoding name using its
+// numeric value and under Unix we do it the other way round, but this just
+// reflects the way different systems provide he encoding info
+
+/* static */
+wxString wxLocale::GetSystemEncodingName()
+{
+    wxString encname;
+
+#ifdef __WIN32__
+    // FIXME: what is the error return value for GetACP()?
+    UINT codepage = ::GetACP();
+    encname.Printf(_T("cp%u"), codepage);
+#elif defined(__UNIX_LIKE__)
+
+#if defined(HAVE_LANGINFO_H) && defined(CODESET)
+    // GNU libc provides current character set this way (this conforms
+    // to Unix98)
+    char *alang = nl_langinfo(CODESET);
+    if (alang)
+    {
+        encname = wxConvLibc.cMB2WX(alang);
+    }
+    else
+#endif // HAVE_LANGINFO_H
+    {
+        // if we can't get at the character set directly, try to see if it's in
+        // the environment variables (in most cases this won't work, but I was
+        // out of ideas)
+        wxChar *lang = wxGetenv(wxT("LC_ALL"));
+        wxChar *dot = lang ? wxStrchr(lang, wxT('.')) : (wxChar *)NULL;
+        if (!dot)
+        {
+            lang = wxGetenv(wxT("LC_CTYPE"));
+            if ( lang )
+                dot = wxStrchr(lang, wxT('.'));
+        }
+        if (!dot)
+        {
+            lang = wxGetenv(wxT("LANG"));
+            if ( lang )
+                dot = wxStrchr(lang, wxT('.'));
+        }
+
+        if ( dot )
+        {
+            encname = dot+1;
+        }
+    }
+#endif // Win32/Unix
+
+    return encname;
+}
+
+/* static */
+wxFontEncoding wxLocale::GetSystemEncoding()
+{
+#ifdef __WIN32__
+    UINT codepage = ::GetACP();
+
+    // wxWindows only knows about CP1250-1257
+    if ( codepage >= 1250 && codepage <= 1257 )
+    {
+        return (wxFontEncoding)(wxFONTENCODING_CP1250 + codepage - 1250);
+    }
+#elif defined(__UNIX_LIKE__)
+    wxString encname = GetSystemEncodingName();
+    if ( !encname.empty() )
+    {
+        return wxFontMapper::CharsetToEncoding(encname,
+                                               FALSE /* not interactive */);
+    }
+#endif // Win32/Unix
+
+    return wxFONTENCODING_SYSTEM;
+}
+
 /*static*/ void wxLocale::AddLanguage(const wxLanguageInfo& info)
 {
     CreateLanguagesDB();

src/common/strconv.cpp

@@ -700,15 +700,24 @@ static wxCharacterSet *wxGetCharacterSet(const wxChar *name)
     }
 
     if (cset && cset->usable()) return cset;
-    if (cset) delete cset;
+    if (cset)
+    {
+        delete cset;
         cset = NULL;
+    }
+
 #ifdef __WIN32__
     cset = new CP_CharSet(name); // may take NULL
-    if (cset->usable()) return cset;
+    if (cset->usable())
-#endif
+        return cset;
-    if (cset) delete cset;
+
+    delete cset;
+#endif // __WIN32__
+
     cset = new EC_CharSet(name);
-    if (cset->usable()) return cset;
+    if (cset->usable())
+        return cset;
+
     delete cset;
     wxLogError(_("Unknown encoding '%s'!"), name);
     return NULL;
@@ -716,16 +725,15 @@ static wxCharacterSet *wxGetCharacterSet(const wxChar *name)
 
 wxCSConv::wxCSConv(const wxChar *charset)
 {
-    m_name = (wxChar *) NULL;
+    m_name = (wxChar *)NULL;
     m_cset = (wxCharacterSet *) NULL;
-    m_deferred = TRUE;
     SetName(charset);
 }
 
 wxCSConv::~wxCSConv()
 {
-    if (m_name) free(m_name);
+    free(m_name);
-    if (m_cset) delete m_cset;
+    delete m_cset;
 }
 
 void wxCSConv::SetName(const wxChar *charset)
@@ -741,39 +749,13 @@ void wxCSConv::LoadNow()
 {
     if (m_deferred)
     {
-        if (!m_name)
+        if ( !m_name )
         {
-#ifdef __UNIX__
+            wxString name = wxLocale::GetSystemEncodingName();
-#if defined(HAVE_LANGINFO_H) && defined(CODESET)
+            if ( !name.empty() )
-            // GNU libc provides current character set this way
+                SetName(name);
-            char *alang = nl_langinfo(CODESET);
-            if (alang)
-            {
-                SetName(wxConvLibc.cMB2WX(alang));
-            }
-            else
-#endif
-            {
-                // if we can't get at the character set directly,
-                // try to see if it's in the environment variables
-                // (in most cases this won't work, but I was out of ideas)
-                wxChar *lang = wxGetenv(wxT("LC_ALL"));
-                wxChar *dot = lang ? wxStrchr(lang, wxT('.')) : (wxChar *)NULL;
-                if (!dot)
-                {
-                    lang = wxGetenv(wxT("LC_CTYPE"));
-                    dot = lang ? wxStrchr(lang, wxT('.')) : (wxChar *)NULL;
-                }
-                if (!dot)
-                {
-                    lang = wxGetenv(wxT("LANG"));
-                    dot = lang ? wxStrchr(lang, wxT('.')) : (wxChar *)NULL;
-                }
-                if (dot)
-                    SetName(dot+1);
-            }
-#endif
         }
+
         m_cset = wxGetCharacterSet(m_name);
         m_deferred = FALSE;
     }
@@ -821,6 +803,7 @@ size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const
 }
 
 #ifdef HAVE_ICONV_H
+
 class IC_CharSetConverter
 {
 public:
@@ -853,7 +836,8 @@ public:
 public:
     iconv_t cnv;
 };
-#endif
+
+#endif // HAVE_ICONV_H
 
 class EC_CharSetConverter
 {