Added includes to correct compilo

[wxWidgets.git] / src / common / strconv.cpp

diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp
index 8b97f59337961040e0657082aa748dd4f633697b..c8207ca44e4fa48f52ecd9ecb4942c7fdc323565 100644 (file)
--- a/src/common/strconv.cpp
+++ b/src/common/strconv.cpp
@@ -28,8 +28,13 @@
   #pragma hdrstop
 #endif
 
   #pragma hdrstop
 #endif
 

+#ifndef WX_PRECOMP
+    #include "wx/intl.h"
+    #include "wx/log.h"
+#endif // WX_PRECOMP
+

 #ifdef __WXMSW__
 #ifdef __WXMSW__

-  #include "wx/msw/private.h"
+    #include "wx/msw/private.h"

 #endif
 
 #include <errno.h>
 #endif
 
 #include <errno.h>


@@ -37,24 +42,43 @@
 #include <string.h>
 #include <stdlib.h>
 
 #include <string.h>
 #include <stdlib.h>
 

+#include "wx/strconv.h"
+
+// ----------------------------------------------------------------------------
+// globals
+// ----------------------------------------------------------------------------
+
+#if wxUSE_WCHAR_T
+    WXDLLEXPORT_DATA(wxMBConv) wxConvLibc;
+    WXDLLEXPORT_DATA(wxCSConv) wxConvLocal((const wxChar *)NULL);
+#else
+    // stand-ins in absence of wchar_t
+    WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile;
+#endif // wxUSE_WCHAR_T
+
+WXDLLEXPORT_DATA(wxMBConv *) wxConvCurrent = &wxConvLibc;
+
+// ----------------------------------------------------------------------------
+// headers
+// ----------------------------------------------------------------------------
+
+#if wxUSE_WCHAR_T
+

 #ifdef __SALFORDC__
 #ifdef __SALFORDC__

-  #include <clib.h>
+    #include <clib.h>

 #endif
 
 #endif
 

-#ifdef HAVE_ICONV_H
-  #include <iconv.h>
+#ifdef HAVE_ICONV
+    #include <iconv.h>

 #endif
 
 #endif
 

-#ifdef __WXMSW__
-  #include <windows.h>
-#endif
+#include "wx/encconv.h"
+#include "wx/fontmap.h"

 
 

-#include "wx/debug.h"
-#include "wx/strconv.h"
-#include "wx/intl.h"
-#include "wx/log.h"
+// ----------------------------------------------------------------------------
+// macros
+// ----------------------------------------------------------------------------

 
 

-#if wxUSE_WCHAR_T

 #define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT32_SWAP_ALWAYS(str[_c]); }
 #define BSWAP_UTF16(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT16_SWAP_ALWAYS(str[_c]); }
 
 #define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT32_SWAP_ALWAYS(str[_c]); }
 #define BSWAP_UTF16(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT16_SWAP_ALWAYS(str[_c]); }
 


@@ -86,23 +110,21 @@
     #error "Weird sizeof(wchar_t): please report your platform details to wx-users mailing list"
 #endif
 
     #error "Weird sizeof(wchar_t): please report your platform details to wx-users mailing list"
 #endif
 

-// ----------------------------------------------------------------------------
-// globals
-// ----------------------------------------------------------------------------
-
-WXDLLEXPORT_DATA(wxMBConv *) wxConvCurrent = &wxConvLibc;
-

 // ============================================================================
 // implementation
 // ============================================================================
 
 // ============================================================================
 // implementation
 // ============================================================================
 

+// ----------------------------------------------------------------------------
+// UTF-16 en/decoding
+// ----------------------------------------------------------------------------
+

 #ifdef WC_UTF16
 
 static size_t encode_utf16(wxUint32 input, wchar_t *output)
 {
     if (input<=0xffff)
     {
 #ifdef WC_UTF16
 
 static size_t encode_utf16(wxUint32 input, wchar_t *output)
 {
     if (input<=0xffff)
     {

-        if (output) *output++ = input;
+        if (output) *output++ = (wchar_t) input;

         return 1;
     }
     else if (input>=0x110000)
         return 1;
     }
     else if (input>=0x110000)


@@ -113,8 +135,8 @@ static size_t encode_utf16(wxUint32 input, wchar_t *output)
     {
         if (output)
         {
     {
         if (output)
         {

-            *output++ = (input >> 10)+0xd7c0;
-            *output++ = (input&0x3ff)+0xdc00;
+            *output++ = (wchar_t) ((input >> 10)+0xd7c0);
+            *output++ = (wchar_t) ((input&0x3ff)+0xdc00);

         }
         return 2;
     }
         }
         return 2;
     }


@@ -145,8 +167,6 @@ static size_t decode_utf16(const wchar_t* input, wxUint32& output)
 // wxMBConv
 // ----------------------------------------------------------------------------
 
 // wxMBConv
 // ----------------------------------------------------------------------------
 

-WXDLLEXPORT_DATA(wxMBConv) wxConvLibc;
-

 size_t wxMBConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const
 {
     return wxMB2WC(buf, psz, n);
 size_t wxMBConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const
 {
     return wxMB2WC(buf, psz, n);


@@ -337,11 +357,11 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const
                 if (buf)
                     buf += pa;
                 len += pa;
                 if (buf)
                     buf += pa;
                 len += pa;

-#else
+#else // !WC_UTF16

                 if (buf)
                     *buf++ = res;
                 len++;
                 if (buf)
                     *buf++ = res;
                 len++;

-#endif
+#endif // WC_UTF16/!WC_UTF16

             }
         }
     }
             }
         }
     }


@@ -369,7 +389,7 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const
         {
             // plain ASCII char
             if (buf)
         {
             // plain ASCII char
             if (buf)

-                *buf++ = cc;
+                *buf++ = (char) cc;

             len++;
         }
 
             len++;
         }
 


@@ -378,9 +398,9 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const
             len += cnt + 1;
             if (buf)
             {
             len += cnt + 1;
             if (buf)
             {

-                *buf++ = (-128 >> cnt) | ((cc >> (cnt * 6)) & (0x3f >> cnt));
+                *buf++ = (char) ((-128 >> cnt) | ((cc >> (cnt * 6)) & (0x3f >> cnt)));

                 while (cnt--)
                 while (cnt--)

-                    *buf++ = 0x80 | ((cc >> (cnt * 6)) & 0x3f);
+                    *buf++ = (char) (0x80 | ((cc >> (cnt * 6)) & 0x3f));

             }
         }
     }
             }
         }
     }


@@ -389,76 +409,34 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const
     return len;
 }
 
     return len;
 }
 

+// ============================================================================
+// wxCharacterSet and derived classes
+// ============================================================================
+

 // ----------------------------------------------------------------------------
 // ----------------------------------------------------------------------------

-// specified character set
+// wxCharacterSet is the ABC for the classes below

 // ----------------------------------------------------------------------------
 
 // ----------------------------------------------------------------------------
 

-WXDLLEXPORT_DATA(wxCSConv) wxConvLocal((const wxChar *)NULL);
-
-#include "wx/encconv.h"
-#include "wx/fontmap.h"
-
-// TODO: add some tables here
-// - perhaps common encodings to common codepages (for Win32)
-// - perhaps common encodings to objects ("UTF8" -> wxConvUTF8)
-// - move wxEncodingConverter meat in here
-
-#if defined(__WIN32__) && !defined(__WXMICROWIN__)
-#include "wx/msw/registry.h"
-// this should work if M$ Internet Exploiter is installed
-static long CharsetToCodepage(const wxChar *name)
-{
-    if (!name)
-        return GetACP();
-
-    long CP=-1;
-
-    wxString cn(name);
-    do {
-        wxString path(wxT("MIME\\Database\\Charset\\"));
-        path += cn;
-        wxRegKey key(wxRegKey::HKCR, path);
-
-        if (!key.Exists()) break;
-
-        // two cases: either there's an AliasForCharset string,
-        // or there are Codepage and InternetEncoding dwords.
-        // The InternetEncoding gives us the actual encoding,
-        // the Codepage just says which Windows character set to
-        // use when displaying the data.
-        if (key.HasValue(wxT("InternetEncoding")) &&
-            key.QueryValue(wxT("InternetEncoding"), &CP)) break;
-
-        // no encoding, see if it's an alias
-        if (!key.HasValue(wxT("AliasForCharset")) ||
-            !key.QueryValue(wxT("AliasForCharset"), cn)) break;
-    } while (1);
-
-    return CP;
-}
-#endif
-

 class wxCharacterSet
 {
 public:
 class wxCharacterSet
 {
 public:

-    wxCharacterSet(const wxChar*name)
-        : cname(name) {}
-    virtual ~wxCharacterSet()
-        {}
-    virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n)
-        { return (size_t)-1; }
-    virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n)
-        { return (size_t)-1; }
-    virtual bool usable()
-        { return FALSE; }
+    wxCharacterSet(const wxChar*name) : cname(name) {}
+    virtual ~wxCharacterSet() {}
+    virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) = 0;
+    virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) = 0;
+    virtual bool usable() const = 0;

 public:
     const wxChar*cname;
 };
 
 public:
     const wxChar*cname;
 };
 

+// ----------------------------------------------------------------------------
+// ID_CharSet: implementation of wxCharacterSet using an existing wxMBConv
+// ----------------------------------------------------------------------------
+

 class ID_CharSet : public wxCharacterSet
 {
 public:
 class ID_CharSet : public wxCharacterSet
 {
 public:

-    ID_CharSet(const wxChar *name,wxMBConv *cnv)
+    ID_CharSet(const wxChar *name, wxMBConv *cnv)

         : wxCharacterSet(name), work(cnv) {}
 
     size_t MB2WC(wchar_t *buf, const char *psz, size_t n)
         : wxCharacterSet(name), work(cnv) {}
 
     size_t MB2WC(wchar_t *buf, const char *psz, size_t n)


@@ -467,17 +445,18 @@ public:
     size_t WC2MB(char *buf, const wchar_t *psz, size_t n)
         { return work ? work->WC2MB(buf,psz,n) : (size_t)-1; }
 
     size_t WC2MB(char *buf, const wchar_t *psz, size_t n)
         { return work ? work->WC2MB(buf,psz,n) : (size_t)-1; }
 

-    bool usable()
+    bool usable() const

         { return work!=NULL; }
 public:
     wxMBConv*work;
 };
 
 
         { return work!=NULL; }
 public:
     wxMBConv*work;
 };
 
 

-#ifdef HAVE_ICONV_H
+// ============================================================================
+// The classes doing conversion using the iconv_xxx() functions
+// ============================================================================

 
 

-bool g_wcNeedsSwap = FALSE;
-static const char *g_wcCharset = NULL;
+#ifdef HAVE_ICONV

 
 // VS: glibc 2.1.3 is broken in that iconv() conversion to/from UCS4 fails with E2BIG
 //     if output buffer is _exactly_ as big as needed. Such case is (unless there's
 
 // VS: glibc 2.1.3 is broken in that iconv() conversion to/from UCS4 fails with E2BIG
 //     if output buffer is _exactly_ as big as needed. Such case is (unless there's


@@ -493,225 +472,269 @@ static const char *g_wcCharset = NULL;
 #define ICONV_FAILED(cres, bufLeft)  (cres == (size_t)-1)
 #endif
 
 #define ICONV_FAILED(cres, bufLeft)  (cres == (size_t)-1)
 #endif
 

+#define ICONV_CHAR_CAST(x)  ((ICONV_CONST char **)(x))
+
+// ----------------------------------------------------------------------------
+// IC_CharSet: encapsulates an iconv character set
+// ----------------------------------------------------------------------------
+

 class IC_CharSet : public wxCharacterSet
 {
 public:
 class IC_CharSet : public wxCharacterSet
 {
 public:

-    IC_CharSet(const wxChar *name)
-        : wxCharacterSet(name)
+    IC_CharSet(const wxChar *name);
+    virtual ~IC_CharSet();
+
+    virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n);
+    virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n);
+
+    bool usable() const
+        { return (m2w != (iconv_t)-1) && (w2m != (iconv_t)-1); }
+
+protected:
+    // the iconv handlers used to translate from multibyte to wide char and in
+    // the other direction
+    iconv_t m2w,
+            w2m;
+
+private:
+    // the name (for iconv_open()) of a wide char charset - if none is
+    // available on this machine, it will remain NULL
+    static const char *ms_wcCharsetName;
+
+    // true if the wide char encoding we use (i.e. ms_wcCharsetName) has
+    // different endian-ness than the native one
+    static bool ms_wcNeedsSwap;
+};
+
+const char *IC_CharSet::ms_wcCharsetName = NULL;
+bool IC_CharSet::ms_wcNeedsSwap = FALSE;
+
+IC_CharSet::IC_CharSet(const wxChar *name)
+          : wxCharacterSet(name)
+{
+    // check for charset that represents wchar_t:
+    if (ms_wcCharsetName == NULL)

     {
     {

-        // check for charset that represents wchar_t:
-        if (g_wcCharset == NULL)
-        {
-            g_wcNeedsSwap = FALSE;
+        ms_wcNeedsSwap = FALSE;
+
+        // try charset with explicit bytesex info (e.g. "UCS-4LE"):
+        ms_wcCharsetName = WC_NAME_BEST;
+        m2w = iconv_open(ms_wcCharsetName, wxConvLibc.cWX2MB(name));

 
 

-            // try charset with explicit bytesex info (e.g. "UCS-4LE"):
-            g_wcCharset = WC_NAME_BEST;
-            m2w = iconv_open(g_wcCharset, wxConvLibc.cWX2MB(name));
+        if (m2w == (iconv_t)-1)
+        {
+            // try charset w/o bytesex info (e.g. "UCS4")
+            // and check for bytesex ourselves:
+            ms_wcCharsetName = WC_NAME;
+            m2w = iconv_open(ms_wcCharsetName, wxConvLibc.cWX2MB(name));

 
 

+            // last bet, try if it knows WCHAR_T pseudo-charset

             if (m2w == (iconv_t)-1)
             {
             if (m2w == (iconv_t)-1)
             {

-                // try charset w/o bytesex info (e.g. "UCS4")
-                // and check for bytesex ourselves: 
-                g_wcCharset = WC_NAME;
-                m2w = iconv_open(g_wcCharset, wxConvLibc.cWX2MB(name));
+                ms_wcCharsetName = "WCHAR_T";
+                m2w = iconv_open(ms_wcCharsetName, wxConvLibc.cWX2MB(name));
+            }

 
 

-                // last bet, try if it knows WCHAR_T pseudo-charset
-                if (m2w == (iconv_t)-1)
+            if (m2w != (iconv_t)-1)
+            {
+                char    buf[2], *bufPtr;
+                wchar_t wbuf[2], *wbufPtr;
+                size_t  insz, outsz;
+                size_t  res;
+
+                buf[0] = 'A';
+                buf[1] = 0;
+                wbuf[0] = 0;
+                insz = 2;
+                outsz = SIZEOF_WCHAR_T * 2;
+                wbufPtr = wbuf;
+                bufPtr = buf;
+
+                res = iconv(m2w, ICONV_CHAR_CAST(&bufPtr), &insz,
+                            (char**)&wbufPtr, &outsz);
+
+                if (ICONV_FAILED(res, insz))

                 {
                 {

-                    g_wcCharset = "WCHAR_T";
-                    m2w = iconv_open(g_wcCharset, wxConvLibc.cWX2MB(name));
-                }
-                
-                if (m2w != (iconv_t)-1)
-                {
-                    char    buf[2], *bufPtr;
-                    wchar_t wbuf[2], *wbufPtr;
-                    size_t  insz, outsz;
-                    size_t  res;
-                    
-                    buf[0] = 'A';
-                    buf[1] = 0;
-                    wbuf[0] = 0;
-                    insz = 2;
-                    outsz = SIZEOF_WCHAR_T * 2;
-                    wbufPtr = wbuf;
-                    bufPtr = buf;
-
-                    #ifdef WX_ICONV_TAKES_CHAR
-                    res = iconv(m2w, (char**)&bufPtr, &insz, (char**)&wbufPtr, &outsz);
-                    #else
-                    res = iconv(m2w, (const char**)&bufPtr, &insz, (char**)&wbufPtr, &outsz);
-                    #endif
-                    if (ICONV_FAILED(res, insz))
-                    {
-                        g_wcCharset = NULL;
-                        wxLogLastError(wxT("iconv"));
-                        wxLogError(_("Convertion to charset '%s' doesn't work."), name);
-                    }
-                    else
-                    {
-                        g_wcNeedsSwap = (wbuf[0] != (wchar_t)buf[0]);
-                    }
+                    ms_wcCharsetName = NULL;
+                    wxLogLastError(wxT("iconv"));
+                    wxLogError(_("Convertion to charset '%s' doesn't work."), name);

                 }
                 else
                 {
                 }
                 else
                 {

-                    g_wcCharset = NULL;
-                    wxLogError(_("Don't know how to convert to/from charset '%s'."), name);
+                    ms_wcNeedsSwap = wbuf[0] != (wchar_t)buf[0];

                 }
             }
                 }
             }

-            wxLogTrace(wxT("strconv"), wxT("wchar_t charset is '%s', needs swap: %i"), g_wcCharset, g_wcNeedsSwap);
+            else
+            {
+                ms_wcCharsetName = NULL;
+
+                // VS: we must not output an error here, since wxWindows will safely
+                //     fall back to using wxEncodingConverter.
+                wxLogTrace(wxT("strconv"), wxT("Impossible to convert to/from charset '%s' with iconv, falling back to wxEncodingConverter."), name);
+                //wxLogError(
+            }

         }
         }

-        else
-            m2w = iconv_open(g_wcCharset, wxConvLibc.cWX2MB(name));
-    
-        w2m = iconv_open(wxConvLibc.cWX2MB(name), g_wcCharset);
+        wxLogTrace(wxT("strconv"), wxT("wchar_t charset is '%s', needs swap: %i"), ms_wcCharsetName, ms_wcNeedsSwap);

     }
     }

-    
-    ~IC_CharSet() 
+    else // we already have ms_wcCharsetName

     {
     {

-        if ( m2w != (iconv_t)-1 )
-            iconv_close(m2w);
-        if ( w2m != (iconv_t)-1 )
-            iconv_close(w2m);
+        m2w = iconv_open(ms_wcCharsetName, wxConvLibc.cWX2MB(name));

     }
 
     }
 

-    size_t MB2WC(wchar_t *buf, const char *psz, size_t n)
+    // NB: don't ever pass NULL to iconv_open(), it may crash!
+    if ( ms_wcCharsetName )

     {
     {

-        size_t inbuf = strlen(psz);
-        size_t outbuf = n * SIZEOF_WCHAR_T;
-        size_t res, cres;
-        // VS: Use these instead of psz, buf because iconv() modifies its arguments:
-        wchar_t *bufPtr = buf;
-        const char *pszPtr = psz;
+        w2m = iconv_open(wxConvLibc.cWX2MB(name), ms_wcCharsetName);
+    }
+    else
+    {
+        w2m = (iconv_t)-1;
+    }
+}

 
 

-        if (buf)
-        {
-            // have destination buffer, convert there
-#ifdef WX_ICONV_TAKES_CHAR
-            cres = iconv(m2w, (char**)&pszPtr, &inbuf, (char**)&bufPtr, &outbuf);
-#else
-            cres = iconv(m2w, &pszPtr, &inbuf, (char**)&bufPtr, &outbuf);
-#endif
-            res = n - (outbuf / SIZEOF_WCHAR_T);
+IC_CharSet::~IC_CharSet()
+{
+    if ( m2w != (iconv_t)-1 )
+        iconv_close(m2w);
+    if ( w2m != (iconv_t)-1 )
+        iconv_close(w2m);
+}

 
 

-            if (g_wcNeedsSwap)
-            {
-                // convert to native endianness
-                WC_BSWAP(buf /* _not_ bufPtr */, res)
-            }
-        }
-        else
-        {
-            // no destination buffer... convert using temp buffer
-            // to calculate destination buffer requirement
-            wchar_t tbuf[8];
-            res = 0;
-            do {
-                bufPtr = tbuf; outbuf = 8*SIZEOF_WCHAR_T;
-#ifdef WX_ICONV_TAKES_CHAR
-                cres = iconv( m2w, (char**)&pszPtr, &inbuf, (char**)&bufPtr, &outbuf );
-#else
-                cres = iconv( m2w, &pszPtr, &inbuf, (char**)&bufPtr, &outbuf );
-#endif
-                res += 8-(outbuf/SIZEOF_WCHAR_T);
-            } while ((cres==(size_t)-1) && (errno==E2BIG));
-        }
+size_t IC_CharSet::MB2WC(wchar_t *buf, const char *psz, size_t n)
+{
+    size_t inbuf = strlen(psz);
+    size_t outbuf = n * SIZEOF_WCHAR_T;
+    size_t res, cres;
+    // VS: Use these instead of psz, buf because iconv() modifies its arguments:
+    wchar_t *bufPtr = buf;
+    const char *pszPtr = psz;

 
 

-        if (ICONV_FAILED(cres, inbuf))
+    if (buf)
+    {
+        // have destination buffer, convert there
+        cres = iconv(m2w,
+                     ICONV_CHAR_CAST(&pszPtr), &inbuf,
+                     (char**)&bufPtr, &outbuf);
+        res = n - (outbuf / SIZEOF_WCHAR_T);
+
+        if (ms_wcNeedsSwap)

         {
         {

-            //VS: it is ok if iconv fails, hence trace only
-            wxLogTrace(wxT("strconv"), wxT("iconv failed: %s"), wxSysErrorMsg(wxSysErrorCode()));
-            return (size_t)-1;
+            // convert to native endianness
+            WC_BSWAP(buf /* _not_ bufPtr */, res)

         }
         }

+    }
+    else
+    {
+        // no destination buffer... convert using temp buffer
+        // to calculate destination buffer requirement
+        wchar_t tbuf[8];
+        res = 0;
+        do {
+            bufPtr = tbuf;
+            outbuf = 8*SIZEOF_WCHAR_T;
+
+            cres = iconv(m2w,
+                         ICONV_CHAR_CAST(&pszPtr), &inbuf,
+                         (char**)&bufPtr, &outbuf );

 
 

-        return res;
+            res += 8-(outbuf/SIZEOF_WCHAR_T);
+        } while ((cres==(size_t)-1) && (errno==E2BIG));

     }
 
     }
 

-    size_t WC2MB(char *buf, const wchar_t *psz, size_t n)
+    if (ICONV_FAILED(cres, inbuf))

     {
     {

+        //VS: it is ok if iconv fails, hence trace only
+        wxLogTrace(wxT("strconv"), wxT("iconv failed: %s"), wxSysErrorMsg(wxSysErrorCode()));
+        return (size_t)-1;
+    }
+
+    return res;
+}
+
+size_t IC_CharSet::WC2MB(char *buf, const wchar_t *psz, size_t n)
+{

 #if defined(__BORLANDC__) && (__BORLANDC__ > 0x530)
 #if defined(__BORLANDC__) && (__BORLANDC__ > 0x530)

-        size_t inbuf = std::wcslen(psz) * SIZEOF_WCHAR_T;
+    size_t inbuf = std::wcslen(psz) * SIZEOF_WCHAR_T;

 #else
 #else

-        size_t inbuf = ::wcslen(psz) * SIZEOF_WCHAR_T;
+    size_t inbuf = ::wcslen(psz) * SIZEOF_WCHAR_T;

 #endif
 #endif

-        size_t outbuf = n;
-        size_t res, cres;
-        
-        wchar_t *tmpbuf;
+    size_t outbuf = n;
+    size_t res, cres;

 
 

-        if (g_wcNeedsSwap)
-        {
-            // need to copy to temp buffer to switch endianness
-            // this absolutely doesn't rock!
-            // (no, doing WC_BSWAP twice on the original buffer won't help, as it
-            //  could be in read-only memory, or be accessed in some other thread)
-            tmpbuf=(wchar_t*)malloc((inbuf+1)*SIZEOF_WCHAR_T);
-            memcpy(tmpbuf,psz,(inbuf+1)*SIZEOF_WCHAR_T);
-            WC_BSWAP(tmpbuf, inbuf)
-            psz=tmpbuf;
-        }
+    wchar_t *tmpbuf = 0;

 
 

-        if (buf)
-        {
-            // have destination buffer, convert there
-#ifdef WX_ICONV_TAKES_CHAR
-            cres = iconv( w2m, (char**)&psz, &inbuf, &buf, &outbuf );
-#else
-            cres = iconv( w2m, (const char**)&psz, &inbuf, &buf, &outbuf );
-#endif
-            res = n-outbuf;
-        }
-        else
-        {
-            // no destination buffer... convert using temp buffer
-            // to calculate destination buffer requirement
-            char tbuf[16];
-            res = 0;
-            do {
-                buf = tbuf; outbuf = 16;
-#ifdef WX_ICONV_TAKES_CHAR
-                cres = iconv( w2m, (char**)&psz, &inbuf, &buf, &outbuf );
-#else
-                cres = iconv( w2m, (const char**)&psz, &inbuf, &buf, &outbuf );
-#endif
-                res += 16 - outbuf;
-            } while ((cres==(size_t)-1) && (errno==E2BIG));
-        }
+    if (ms_wcNeedsSwap)
+    {
+        // need to copy to temp buffer to switch endianness
+        // this absolutely doesn't rock!
+        // (no, doing WC_BSWAP twice on the original buffer won't help, as it
+        //  could be in read-only memory, or be accessed in some other thread)
+        tmpbuf=(wchar_t*)malloc((inbuf+1)*SIZEOF_WCHAR_T);
+        memcpy(tmpbuf,psz,(inbuf+1)*SIZEOF_WCHAR_T);
+        WC_BSWAP(tmpbuf, inbuf)
+        psz=tmpbuf;
+    }

 
 

-        if (g_wcNeedsSwap)
-        {
-            free(tmpbuf);
-        }
+    if (buf)
+    {
+        // have destination buffer, convert there
+        cres = iconv( w2m, ICONV_CHAR_CAST(&psz), &inbuf, &buf, &outbuf );

 
 

-        if (ICONV_FAILED(cres, inbuf))
-        {
-            //VS: it is ok if iconv fails, hence trace only
-            wxLogTrace(wxT("strconv"), wxT("iconv failed: %s"), wxSysErrorMsg(wxSysErrorCode()));
-            return (size_t)-1;
-        }
+        res = n-outbuf;
+    }
+    else
+    {
+        // no destination buffer... convert using temp buffer
+        // to calculate destination buffer requirement
+        char tbuf[16];
+        res = 0;
+        do {
+            buf = tbuf; outbuf = 16;
+
+            cres = iconv( w2m, ICONV_CHAR_CAST(&psz), &inbuf, &buf, &outbuf );

 
 

-        return res;
+            res += 16 - outbuf;
+        } while ((cres==(size_t)-1) && (errno==E2BIG));

     }
 
     }
 

-    bool usable()
-        { return (m2w != (iconv_t)-1) && (w2m != (iconv_t)-1); }
+    if (ms_wcNeedsSwap)
+    {
+        free(tmpbuf);
+    }

 
 

-protected:
-    iconv_t m2w, w2m;
-};
-#endif
+    if (ICONV_FAILED(cres, inbuf))
+    {
+        //VS: it is ok if iconv fails, hence trace only
+        wxLogTrace(wxT("strconv"), wxT("iconv failed: %s"), wxSysErrorMsg(wxSysErrorCode()));
+        return (size_t)-1;
+    }
+
+    return res;
+}
+
+#endif // HAVE_ICONV
+
+// ============================================================================
+// Win32 conversion classes
+// ============================================================================

 
 #if defined(__WIN32__) && !defined(__WXMICROWIN__)
 
 #if defined(__WIN32__) && !defined(__WXMICROWIN__)

+
+extern long wxCharsetToCodepage(const wxChar *charset); // from utils.cpp
+

 class CP_CharSet : public wxCharacterSet
 {
 public:
 class CP_CharSet : public wxCharacterSet
 {
 public:

-    CP_CharSet(const wxChar*name)
-        : wxCharacterSet(name), CodePage(CharsetToCodepage(name)) {}
+    CP_CharSet(const wxChar* name)
+        : wxCharacterSet(name)
+        {
+            m_CodePage = wxCharsetToCodepage(name);
+        }

 
     size_t MB2WC(wchar_t *buf, const char *psz, size_t n)
     {
         size_t len =
 
     size_t MB2WC(wchar_t *buf, const char *psz, size_t n)
     {
         size_t len =

-            MultiByteToWideChar(CodePage, 0, psz, -1, buf, buf ? n : 0);
+            MultiByteToWideChar(m_CodePage, 0, psz, -1, buf, buf ? n : 0);

         //VS: returns # of written chars for buf!=NULL and *size*
         //    needed buffer for buf==NULL
         return len ? (buf ? len : len-1) : (size_t)-1;
         //VS: returns # of written chars for buf!=NULL and *size*
         //    needed buffer for buf==NULL
         return len ? (buf ? len : len-1) : (size_t)-1;


@@ -719,21 +742,25 @@ public:
 
     size_t WC2MB(char *buf, const wchar_t *psz, size_t n)
     {
 
     size_t WC2MB(char *buf, const wchar_t *psz, size_t n)
     {

-        size_t len = WideCharToMultiByte(CodePage, 0, psz, -1, buf,
+        size_t len = WideCharToMultiByte(m_CodePage, 0, psz, -1, buf,

                                          buf ? n : 0, NULL, NULL);
         //VS: returns # of written chars for buf!=NULL and *size*
         //    needed buffer for buf==NULL
         return len ? (buf ? len : len-1) : (size_t)-1;
     }
 
                                          buf ? n : 0, NULL, NULL);
         //VS: returns # of written chars for buf!=NULL and *size*
         //    needed buffer for buf==NULL
         return len ? (buf ? len : len-1) : (size_t)-1;
     }
 

-    bool usable()
-        { return CodePage != -1; }
+    bool usable() const
+        { return m_CodePage != -1; }

 
 public:
 
 public:

-    long CodePage;
+    long m_CodePage;

 };
 #endif // __WIN32__
 
 };
 #endif // __WIN32__
 

+// ============================================================================
+// wxEncodingConverter based conversion classes
+// ============================================================================
+

 #if wxUSE_FONTMAP
 
 class EC_CharSet : public wxCharacterSet
 #if wxUSE_FONTMAP
 
 class EC_CharSet : public wxCharacterSet


@@ -741,16 +768,17 @@ class EC_CharSet : public wxCharacterSet
 public:
     // temporarily just use wxEncodingConverter stuff,
     // so that it works while a better implementation is built
 public:
     // temporarily just use wxEncodingConverter stuff,
     // so that it works while a better implementation is built

-    EC_CharSet(const wxChar*name) : wxCharacterSet(name),
-                                    enc(wxFONTENCODING_SYSTEM)
+    EC_CharSet(const wxChar* name) : wxCharacterSet(name),
+                                     enc(wxFONTENCODING_SYSTEM)

     {
         if (name)
             enc = wxTheFontMapper->CharsetToEncoding(name, FALSE);
     {
         if (name)
             enc = wxTheFontMapper->CharsetToEncoding(name, FALSE);

-        m2w.Init(enc, wxFONTENCODING_UNICODE);
-        w2m.Init(wxFONTENCODING_UNICODE, enc);
+
+        m_ok = m2w.Init(enc, wxFONTENCODING_UNICODE) &&
+               w2m.Init(wxFONTENCODING_UNICODE, enc);

     }
 
     }
 

-    size_t MB2WC(wchar_t *buf, const char *psz, size_t n)
+    size_t MB2WC(wchar_t *buf, const char *psz, size_t WXUNUSED(n))

     {
         size_t inbuf = strlen(psz);
         if (buf)
     {
         size_t inbuf = strlen(psz);
         if (buf)


@@ -758,9 +786,10 @@ public:
         return inbuf;
     }
 
         return inbuf;
     }
 

-    size_t WC2MB(char *buf, const wchar_t *psz, size_t n)
+    size_t WC2MB(char *buf, const wchar_t *psz, size_t WXUNUSED(n))

     {
     {

-#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530)
+#if ( defined(__BORLANDC__) && (__BORLANDC__ > 0x530) ) \
+    || ( defined(__MWERKS__) && defined(__WXMSW__) )

         size_t inbuf = std::wcslen(psz);
 #else
         size_t inbuf = ::wcslen(psz);
         size_t inbuf = std::wcslen(psz);
 #else
         size_t inbuf = ::wcslen(psz);


@@ -771,61 +800,93 @@ public:
         return inbuf;
     }
 
         return inbuf;
     }
 

-    bool usable()
-        { return (enc!=wxFONTENCODING_SYSTEM) && (enc!=wxFONTENCODING_DEFAULT); }
+    bool usable() const { return m_ok; }

 
 public:
     wxFontEncoding enc;
     wxEncodingConverter m2w, w2m;
 
 public:
     wxFontEncoding enc;
     wxEncodingConverter m2w, w2m;

+
+    // were we initialized successfully?
+    bool m_ok;

 };
 
 #endif // wxUSE_FONTMAP
 
 };
 
 #endif // wxUSE_FONTMAP
 

+// ----------------------------------------------------------------------------
+// the function creating the wxCharacterSet for the specified charset on the
+// current system, trying all possibilities
+// ----------------------------------------------------------------------------
+

 static wxCharacterSet *wxGetCharacterSet(const wxChar *name)
 {
 static wxCharacterSet *wxGetCharacterSet(const wxChar *name)
 {

-    wxCharacterSet *cset = NULL;
-    if (name)
+    // check for the special case of ASCII charset
+#if wxUSE_FONTMAP
+    if ( wxTheFontMapper->CharsetToEncoding(name) == wxFONTENCODING_DEFAULT )
+#else // wxUSE_FONTMAP
+    if ( !name )
+#endif // wxUSE_FONTMAP/!wxUSE_FONTMAP

     {
     {

-        if (wxStricmp(name, wxT("UTF8")) == 0 || wxStricmp(name, wxT("UTF-8")) == 0)
-        {
-            cset = new ID_CharSet(name, &wxConvUTF8);
-        }
-        else
-        {
-#ifdef HAVE_ICONV_H
-            cset = new IC_CharSet(name); // may not take NULL
-#endif
-        }
+        // don't convert at all
+        return NULL;

     }
 
     }
 

-    if (cset && cset->usable())
-        return cset;
+    // the test above must have taken care of this case
+    wxCHECK_MSG( name, NULL, _T("NULL name must be wxFONTENCODING_DEFAULT") );
+
+    wxCharacterSet *cset;
+
+    if ( wxStricmp(name, wxT("UTF8")) == 0 || wxStricmp(name, wxT("UTF-8")) == 0)
+    {
+        cset = new ID_CharSet(name, &wxConvUTF8);
+    }
+    else
+    {
+#ifdef HAVE_ICONV
+        cset = new IC_CharSet(name);
+#else // !HAVE_ICONV
+        cset = NULL;
+#endif // HAVE_ICONV/!HAVE_ICONV
+    }

 
 

-    if (cset)
+    // it can only be NULL in this case
+#ifndef HAVE_ICONV
+    if ( cset )
+#endif // !HAVE_ICONV

     {
     {

+        if ( cset->usable() )
+            return cset;
+

         delete cset;
         cset = NULL;
     }
 
 #if defined(__WIN32__) && !defined(__WXMICROWIN__)
         delete cset;
         cset = NULL;
     }
 
 #if defined(__WIN32__) && !defined(__WXMICROWIN__)

-    cset = new CP_CharSet(name); // may take NULL
-    if (cset->usable())
+    cset = new CP_CharSet(name);
+    if ( cset->usable() )

         return cset;
 
     delete cset;
         return cset;
 
     delete cset;

+    cset = NULL;

 #endif // __WIN32__
 
 #if wxUSE_FONTMAP
     cset = new EC_CharSet(name);
 #endif // __WIN32__
 
 #if wxUSE_FONTMAP
     cset = new EC_CharSet(name);

-    if (cset->usable())
+    if ( cset->usable() )

         return cset;
         return cset;

-#endif // wxUSE_FONTMAP

 
     delete cset;
 
     delete cset;

-    wxLogError(_("Unknown encoding '%s'!"), name);
+    cset = NULL;
+#endif // wxUSE_FONTMAP
+
+    wxLogError(_("Cannot convert from encoding '%s'!"), name);
+

     return NULL;
 }
 
     return NULL;
 }
 

+// ============================================================================
+// wxCSConv implementation
+// ============================================================================
+

 wxCSConv::wxCSConv(const wxChar *charset)
 {
     m_name = (wxChar *)NULL;
 wxCSConv::wxCSConv(const wxChar *charset)
 {
     m_name = (wxChar *)NULL;


@@ -894,7 +955,8 @@ size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const
         return m_cset->WC2MB(buf, psz, n);
 
     // latin-1 (direct)
         return m_cset->WC2MB(buf, psz, n);
 
     // latin-1 (direct)

-#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530)
+#if ( defined(__BORLANDC__) && (__BORLANDC__ > 0x530) ) \
+    || ( defined(__MWERKS__) && defined(__WXMSW__) )

     size_t len=std::wcslen(psz);
 #else
     size_t len=::wcslen(psz);
     size_t len=std::wcslen(psz);
 #else
     size_t len=::wcslen(psz);


@@ -908,69 +970,6 @@ size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const
     return len;
 }
 
     return len;
 }
 

-#ifdef HAVE_ICONV_H
-
-class IC_CharSetConverter
-{
-public:
-    IC_CharSetConverter(IC_CharSet *from, IC_CharSet *to)
-    {
-        cnv = iconv_open(wxConvLibc.cWX2MB(to->cname),
-                         wxConvLibc.cWX2MB(from->cname));
-    }
-
-    ~IC_CharSetConverter()
-    {
-        if (cnv != (iconv_t)-1)
-            iconv_close(cnv);
-    }
-
-    size_t Convert(char *buf, const char *psz, size_t n)
-    {
-        size_t inbuf = strlen(psz);
-        size_t outbuf = n;
-#ifdef WX_ICONV_TAKES_CHAR
-        size_t res = iconv( cnv, (char**)&psz, &inbuf, &buf, &outbuf );
-#else
-        size_t res = iconv( cnv, &psz, &inbuf, &buf, &outbuf );
-#endif
-        if (res == (size_t)-1)
-            return (size_t)-1;
-        return (n - outbuf);
-    }
-
-public:
-    iconv_t cnv;
-};
-
-#endif // HAVE_ICONV_H
-
-class EC_CharSetConverter
-{
-public:
-    EC_CharSetConverter(EC_CharSet* from,EC_CharSet* to)
-        { cnv.Init(from->enc,to->enc); }
-
-    size_t Convert(char* buf, const char* psz, size_t n)
-    {
-        size_t inbuf = strlen(psz);
-        if (buf) cnv.Convert(psz,buf);
-        return inbuf;
-    }
-
-public:
-    wxEncodingConverter cnv;
-};
-
-#else // !wxUSE_WCHAR_T
-
-// ----------------------------------------------------------------------------
-// stand-ins in absence of wchar_t
-// ----------------------------------------------------------------------------
-
-WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile;
-WXDLLEXPORT_DATA(wxMBConv *) wxConvCurrent = &wxConvLibc;
-

 #endif // wxUSE_WCHAR_T
 
 
 #endif // wxUSE_WCHAR_T