// Modified by:
// Created: 29/01/98
// RCS-ID: $Id$
-// Copyright: (c) 1999 Ove Kaaven, Robert Roebling, Vadim Zeitlin, Vaclav Slavik
+// Copyright: (c) 1999 Ove Kaaven, Robert Roebling, Vaclav Slavik
+// (c) 2000-2003 Vadim Zeitlin
// Licence: wxWindows licence
/////////////////////////////////////////////////////////////////////////////
#include "wx/log.h"
#endif // WX_PRECOMP
+#include "wx/strconv.h"
+
+#if wxUSE_WCHAR_T
+
#ifdef __WXMSW__
#include "wx/msw/private.h"
#endif
#include <string.h>
#include <stdlib.h>
-#include "wx/module.h"
-#include "wx/strconv.h"
-
-// ----------------------------------------------------------------------------
-// globals
-// ----------------------------------------------------------------------------
-
-#if wxUSE_WCHAR_T
- WXDLLIMPEXP_DATA_BASE(wxMBConv) wxConvLibc;
- WXDLLIMPEXP_DATA_BASE(wxCSConv) wxConvLocal((const wxChar *)NULL);
- WXDLLIMPEXP_DATA_BASE(wxCSConv) wxConvISO8859_1(_T("iso-8859-1"));
-#else
- // stand-ins in absence of wchar_t
- WXDLLIMPEXP_DATA_BASE(wxMBConv) wxConvLibc,
- wxConvFile,
- wxConvISO8859_1,
- wxConvLocal,
- wxConvUTF8;
-#endif // wxUSE_WCHAR_T
-
-WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent = &wxConvLibc;
-
-class wxStrConvModule: public wxModule
-{
-public:
- wxStrConvModule() : wxModule() { }
- virtual bool OnInit() { return TRUE; }
- virtual void OnExit()
- {
-#if wxUSE_WCHAR_T
- wxConvLocal.Clear();
- wxConvISO8859_1.Clear();
-#endif
- }
-
- DECLARE_DYNAMIC_CLASS(wxStrConvModule)
-};
-
-IMPLEMENT_DYNAMIC_CLASS(wxStrConvModule, wxModule)
-
+#if defined(__WIN32__) && !defined(__WXMICROWIN__)
+ #define wxHAVE_WIN32_MB2WC
+#endif // __WIN32__ but !__WXMICROWIN__
// ----------------------------------------------------------------------------
// headers
// ----------------------------------------------------------------------------
-#if wxUSE_WCHAR_T
-
#ifdef __SALFORDC__
#include <clib.h>
#endif
#include "wx/encconv.h"
#include "wx/fontmap.h"
+#ifdef __WXMAC__
+#include "ATSUnicode.h"
+#include "TextCommon.h"
+#include "TextEncodingConverter.h"
+
+#include "wx/mac/private.h" // includes mac headers
+#endif
// ----------------------------------------------------------------------------
// macros
// ----------------------------------------------------------------------------
#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]); }
-// under Unix SIZEOF_WCHAR_T is defined by configure, but under other platforms
-// it might be not defined - assume the most common value
-#ifndef SIZEOF_WCHAR_T
- #define SIZEOF_WCHAR_T 2
-#endif // !defined(SIZEOF_WCHAR_T)
-
#if SIZEOF_WCHAR_T == 4
#define WC_NAME "UCS4"
#define WC_BSWAP BSWAP_UCS4
#define WC_NAME_BEST "UTF-16LE"
#endif
#else // sizeof(wchar_t) != 2 nor 4
- // I don't know what to do about this
- #error "Weird sizeof(wchar_t): please report your platform details to wx-users mailing list"
+ // does this ever happen?
+ #error "Unknown sizeof(wchar_t): please report this to wx-dev@lists.wxwindows.org"
#endif
// ============================================================================
{
if (input<=0xffff)
{
- if (output) *output++ = (wxUint16) input;
+ if (output)
+ *output = (wxUint16) input;
return 1;
}
else if (input>=0x110000)
if (output)
{
*output++ = (wxUint16) ((input >> 10)+0xd7c0);
- *output++ = (wxUint16) ((input&0x3ff)+0xdc00);
+ *output = (wxUint16) ((input&0x3ff)+0xdc00);
}
return 2;
}
// wxMBConv
// ----------------------------------------------------------------------------
-#define IGNORE_LIBC 0
-
wxMBConv::~wxMBConv()
{
// nothing to do here
}
-size_t wxMBConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const
-{
-#if IGNORE_LIBC
- if (buf)
- {
- for (size_t i = 0; i < strlen( psz )+1; i++)
- buf[i] = (wchar_t) psz[i];
- return strlen( psz );
- }
- else
- {
- return strlen( psz );
- }
-#else
- return wxMB2WC(buf, psz, n);
-#endif
-}
-
-size_t wxMBConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const
-{
-#if IGNORE_LIBC
- if (buf)
- {
- for (size_t i = 0; i < wxStrlen( psz )+1; i++)
- buf[i] = (char) psz[i];
- return wxStrlen( psz );
- }
- else
- {
- return wxStrlen( psz );
- }
-#else
- return wxWC2MB(buf, psz, n);
-#endif
-}
-
const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const
{
if ( psz )
}
// ----------------------------------------------------------------------------
-// UTF-7
+// wxMBConvLibc
// ----------------------------------------------------------------------------
-WXDLLIMPEXP_DATA_BASE(wxMBConvUTF7) wxConvUTF7;
+size_t wxMBConvLibc::MB2WC(wchar_t *buf, const char *psz, size_t n) const
+{
+ return wxMB2WC(buf, psz, n);
+}
+
+size_t wxMBConvLibc::WC2MB(char *buf, const wchar_t *psz, size_t n) const
+{
+ return wxWC2MB(buf, psz, n);
+}
+
+// ----------------------------------------------------------------------------
+// UTF-7
+// ----------------------------------------------------------------------------
#if 0
static char utf7_setD[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ"
// UTF-8
// ----------------------------------------------------------------------------
-WXDLLIMPEXP_DATA_BASE(wxMBConvUTF8) wxConvUTF8;
-
static wxUint32 utf8_max[]=
{ 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff, 0xffffffff };
return (size_t)-1;
}
#ifdef WC_UTF16
- size_t pa = encode_utf16(res, buf);
+ // cast is ok because wchar_t == wxUuint16 if WC_UTF16
+ size_t pa = encode_utf16(res, (wxUint16 *)buf);
if (pa == (size_t)-1)
return (size_t)-1;
if (buf)
{
wxUint32 cc;
#ifdef WC_UTF16
- size_t pa = decode_utf16(psz, cc);
+ // cast is ok for WC_UTF16
+ size_t pa = decode_utf16((const wxUint16 *)psz, cc);
psz += (pa == (size_t)-1) ? 1 : pa;
#else
cc=(*psz++) & 0x7fffffff;
// ----------------------------------------------------------------------------
#ifdef WORDS_BIGENDIAN
-#define wxMBConvUTF16straight wxMBConvUTF16BE
-#define wxMBConvUTF16swap wxMBConvUTF16LE
+ #define wxMBConvUTF16straight wxMBConvUTF16BE
+ #define wxMBConvUTF16swap wxMBConvUTF16LE
#else
-#define wxMBConvUTF16swap wxMBConvUTF16BE
-#define wxMBConvUTF16straight wxMBConvUTF16LE
+ #define wxMBConvUTF16swap wxMBConvUTF16BE
+ #define wxMBConvUTF16straight wxMBConvUTF16LE
#endif
-WXDLLIMPEXP_DATA_BASE(wxMBConvUTF16LE) wxConvUTF16LE;
-WXDLLIMPEXP_DATA_BASE(wxMBConvUTF16BE) wxConvUTF16BE;
-
-
-
-
-
#ifdef WC_UTF16
-
// copy 16bit MB to 16bit String
size_t wxMBConvUTF16straight::MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
if (buf)
{
*(wxUint16*)buf = cc[0];
- buf += sizeof(wxUint16);
+ buf += sizeof(wxUint16);
if (pa > 1)
{
*(wxUint16*)buf = cc[1];
{
wxUint32 cc;
- size_t pa=decode_utf16(psz, cc);
+ // cast is ok for WC_UTF16
+ size_t pa = decode_utf16((const wxUint16 *)psz, cc);
if (pa == (size_t)-1)
return pa;
len += sizeof(wxUint32);
psz += pa;
}
- if (buf && len<=n-sizeof(wxUint32)) *(wxUint32*)buf=0;
+
+ if (buf && len<=n-sizeof(wxUint32))
+ *(wxUint32*)buf=0;
return len;
}
len += pa;
psz += sizeof(wxUint32);
}
- if (buf && len<n) *buf=0;
+
+ if (buf && len<n)
+ *buf=0;
return len;
}
{
char cc[4];
- size_t pa=decode_utf16(psz, *(wxUint32*)cc);
+ // cast is ok for WC_UTF16
+ size_t pa=decode_utf16((const wxUint16 *)psz, *(wxUint32*)cc);
if (pa == (size_t)-1)
return pa;
len += sizeof(wxUint32);
psz += pa;
}
- if (buf && len<=n-sizeof(wxUint32)) *(wxUint32*)buf=0;
+
+ if (buf && len<=n-sizeof(wxUint32))
+ *(wxUint32*)buf=0;
return len;
}
len++;
psz += sizeof(wxUint32);
}
- if (buf && len<n) *buf=0;
+
+ if (buf && len<n)
+ *buf=0;
return len;
}
psz++;
}
- if (buf && len<=n-sizeof(wxUint32)) *(wxUint32*)buf=0;
+ if (buf && len<=n-sizeof(wxUint32))
+ *(wxUint32*)buf=0;
return len;
}
len++;
psz += sizeof(wxUint32);
}
- if (buf && len<n) *buf=0;
+
+ if (buf && len<n)
+ *buf=0;
return len;
}
len += sizeof(wxUint32);
psz++;
}
- if (buf && len<=n-sizeof(wxUint32)) *(wxUint32*)buf=0;
+
+ if (buf && len<=n-sizeof(wxUint32))
+ *(wxUint32*)buf=0;
return len;
}
#endif // WC_UTF16
-// ============================================================================
-// wxCharacterSet and derived classes
-// ============================================================================
-
-// ----------------------------------------------------------------------------
-// wxCharacterSet is the ABC for the classes below
-// ----------------------------------------------------------------------------
-
-class wxCharacterSet
-{
-public:
- wxCharacterSet() { }
- 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;
-};
-
-// ----------------------------------------------------------------------------
-// ID_CharSet: implementation of wxCharacterSet using an existing wxMBConv
-// ----------------------------------------------------------------------------
-
-class ID_CharSet : public wxCharacterSet
-{
-public:
- ID_CharSet(wxMBConv *cnv) : work(cnv) {}
-
- size_t MB2WC(wchar_t *buf, const char *psz, size_t n)
- { return work ? work->MB2WC(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() const
- { return work!=NULL; }
-public:
- wxMBConv*work;
-};
-
-
// ============================================================================
// The classes doing conversion using the iconv_xxx() functions
// ============================================================================
#define ICONV_CHAR_CAST(x) ((ICONV_CONST char **)(x))
// ----------------------------------------------------------------------------
-// IC_CharSet: encapsulates an iconv character set
+// wxMBConv_iconv: encapsulates an iconv character set
// ----------------------------------------------------------------------------
-class IC_CharSet : public wxCharacterSet
+class wxMBConv_iconv : public wxMBConv
{
public:
- IC_CharSet(const wxChar *name);
- virtual ~IC_CharSet();
+ wxMBConv_iconv(const wxChar *name);
+ virtual ~wxMBConv_iconv();
- 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);
+ virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const;
+ virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const;
- bool usable() const
+ bool IsOk() const
{ return (m2w != (iconv_t)-1) && (w2m != (iconv_t)-1); }
protected:
w2m;
private:
- // the name (for iconv_open()) of a wide char charset - if none is
+ // 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;
static bool ms_wcNeedsSwap;
};
-const char *IC_CharSet::ms_wcCharsetName = NULL;
-bool IC_CharSet::ms_wcNeedsSwap = FALSE;
+const char *wxMBConv_iconv::ms_wcCharsetName = NULL;
+bool wxMBConv_iconv::ms_wcNeedsSwap = false;
-IC_CharSet::IC_CharSet(const wxChar *name)
+wxMBConv_iconv::wxMBConv_iconv(const wxChar *name)
{
// Do it the hard way
char cname[100];
// check for charset that represents wchar_t:
if (ms_wcCharsetName == NULL)
{
- ms_wcNeedsSwap = FALSE;
+ ms_wcNeedsSwap = false;
// try charset with explicit bytesex info (e.g. "UCS-4LE"):
ms_wcCharsetName = WC_NAME_BEST;
}
}
-IC_CharSet::~IC_CharSet()
+wxMBConv_iconv::~wxMBConv_iconv()
{
if ( m2w != (iconv_t)-1 )
iconv_close(m2w);
iconv_close(w2m);
}
-size_t IC_CharSet::MB2WC(wchar_t *buf, const char *psz, size_t n)
+size_t wxMBConv_iconv::MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
size_t inbuf = strlen(psz);
size_t outbuf = n * SIZEOF_WCHAR_T;
return res;
}
-size_t IC_CharSet::WC2MB(char *buf, const wchar_t *psz, size_t n)
+size_t wxMBConv_iconv::WC2MB(char *buf, const wchar_t *psz, size_t n) const
{
size_t inbuf = wxWcslen(psz) * SIZEOF_WCHAR_T;
size_t outbuf = n;
#endif // HAVE_ICONV
+
// ============================================================================
// Win32 conversion classes
// ============================================================================
-#if defined(__WIN32__) && !defined(__WXMICROWIN__) && !defined(__WXUNIVERSAL__)
+#ifdef wxHAVE_WIN32_MB2WC
// from utils.cpp
extern WXDLLIMPEXP_BASE long wxCharsetToCodepage(const wxChar *charset);
extern WXDLLIMPEXP_BASE long wxEncodingToCodepage(wxFontEncoding encoding);
-class CP_CharSet : public wxCharacterSet
+class wxMBConv_win32 : public wxMBConv
{
public:
- CP_CharSet(const wxChar* name)
- {
- m_CodePage = wxCharsetToCodepage(name);
- }
+ wxMBConv_win32()
+ {
+ m_CodePage = CP_ACP;
+ }
- CP_CharSet(wxFontEncoding encoding)
- {
- m_CodePage = wxEncodingToCodepage(encoding);
- }
+ wxMBConv_win32(const wxChar* name)
+ {
+ m_CodePage = wxCharsetToCodepage(name);
+ }
- size_t MB2WC(wchar_t *buf, const char *psz, size_t n)
+ wxMBConv_win32(wxFontEncoding encoding)
+ {
+ m_CodePage = wxEncodingToCodepage(encoding);
+ }
+
+ size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
const size_t len = ::MultiByteToWideChar
(
buf ? n : 0 // size of output buffer
);
- // note that it returns # of written chars for buf != NULL and *size*
- // of the needed buffer for buf == NULL
- return len ? (buf ? len : len - 1) : (size_t)-1;
+ // note that it returns count of written chars for buf != NULL and size
+ // of the needed buffer for buf == NULL so in either case the length of
+ // the string (which never includes the terminating NUL) is one less
+ return len ? len - 1 : (size_t)-1;
}
- size_t WC2MB(char *buf, const wchar_t *psz, size_t n)
+ size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const
{
const size_t len = ::WideCharToMultiByte
(
NULL // [out] was it used?
);
- // see the comment above!
- return len ? (buf ? len : len - 1) : (size_t)-1;
+ // see the comment above for the reason of "len - 1"
+ return len ? len - 1 : (size_t)-1;
}
- bool usable() const
+ bool IsOk() const
{ return m_CodePage != -1; }
public:
long m_CodePage;
};
-#endif // defined(__WIN32__) && !defined(__WXMICROWIN__) && !defined(__WXUNIVERSAL__)
+
+#endif // wxHAVE_WIN32_MB2WC
+
+// ============================================================================
+// Mac conversion classes
+// ============================================================================
+
+#if defined(__WXMAC__) && defined(TARGET_CARBON)
+
+class wxMBConv_mac : public wxMBConv
+{
+public:
+ wxMBConv_mac()
+ {
+ Init(CFStringGetSystemEncoding()) ;
+ }
+
+ wxMBConv_mac(const wxChar* name)
+ {
+ Init( wxMacGetSystemEncFromFontEnc(wxFontMapper::Get()->CharsetToEncoding(name, FALSE) ) ) ;
+ }
+
+ wxMBConv_mac(wxFontEncoding encoding)
+ {
+ Init( wxMacGetSystemEncFromFontEnc(encoding) );
+ }
+
+ ~wxMBConv_mac()
+ {
+ OSStatus status = noErr ;
+ status = TECDisposeConverter(m_MB2WC_converter);
+ status = TECDisposeConverter(m_WC2MB_converter);
+ }
+
+
+ void Init( TextEncodingBase encoding)
+ {
+ OSStatus status = noErr ;
+ m_char_encoding = encoding ;
+#if SIZEOF_WCHAR_T == 4
+ m_unicode_encoding = CreateTextEncoding(kTextEncodingUnicodeDefault,0,kUnicode32BitFormat) ;
+#else
+ m_unicode_encoding = CreateTextEncoding(kTextEncodingUnicodeDefault,0,kUnicode16BitFormat) ;
+#endif
+ status = TECCreateConverter(&m_MB2WC_converter,
+ m_char_encoding,
+ m_unicode_encoding);
+ status = TECCreateConverter(&m_WC2MB_converter,
+ m_unicode_encoding,
+ m_char_encoding);
+ }
+
+ size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const
+ {
+ OSStatus status = noErr ;
+ ByteCount byteOutLen ;
+ ByteCount byteInLen = strlen(psz) ;
+ ByteCount byteBufferLen = n ;
+ wchar_t *tbuf = NULL ;
+
+ if (buf == NULL)
+ {
+ n = byteInLen * SIZEOF_WCHAR_T ;
+ tbuf = (wchar_t*) malloc( n ) ;
+ }
+
+ status = TECConvertText(m_MB2WC_converter, (ConstTextPtr) psz , byteInLen, &byteInLen,
+ (TextPtr) (buf ? buf : tbuf) , byteBufferLen, &byteOutLen);
+
+ if ( buf == NULL )
+ free(tbuf) ;
+
+ size_t res = byteOutLen / SIZEOF_WCHAR_T ;
+ if ( buf && res < n)
+ buf[res] = 0;
+
+ return res ;
+ }
+
+ size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const
+ {
+ OSStatus status = noErr ;
+ ByteCount byteOutLen ;
+ ByteCount byteInLen = wxWcslen(psz) * SIZEOF_WCHAR_T ;
+ ByteCount byteBufferLen = n ;
+
+ char *tbuf = NULL ;
+
+ if (buf == NULL)
+ {
+ n = byteInLen ;
+ tbuf = (char*) malloc( n ) ;
+ }
+
+ status = TECConvertText(m_WC2MB_converter, (ConstTextPtr) psz , byteInLen, &byteInLen,
+ (TextPtr) ( buf ? buf : tbuf ) , byteBufferLen, &byteOutLen);
+
+ if ( buf == NULL )
+ free(tbuf) ;
+
+ size_t res = byteOutLen ;
+ if ( buf && res < n)
+ buf[res] = 0;
+
+ return res ;
+ }
+
+ bool IsOk() const
+ { return m_MB2WC_converter != NULL && m_WC2MB_converter != NULL ; }
+
+private:
+ TECObjectRef m_MB2WC_converter ;
+ TECObjectRef m_WC2MB_converter ;
+
+ TextEncodingBase m_char_encoding ;
+ TextEncodingBase m_unicode_encoding ;
+};
+
+#endif // defined(__WXMAC__) && defined(TARGET_CARBON)
// ============================================================================
// wxEncodingConverter based conversion classes
#if wxUSE_FONTMAP
-class EC_CharSet : public wxCharacterSet
+class wxMBConv_wxwin : public wxMBConv
{
private:
void Init()
public:
// temporarily just use wxEncodingConverter stuff,
// so that it works while a better implementation is built
- EC_CharSet(const wxChar* name)
+ wxMBConv_wxwin(const wxChar* name)
{
if (name)
- m_enc = wxFontMapper::Get()->CharsetToEncoding(name, FALSE);
+ m_enc = wxFontMapper::Get()->CharsetToEncoding(name, false);
else
m_enc = wxFONTENCODING_SYSTEM;
Init();
}
- EC_CharSet(wxFontEncoding enc)
+ wxMBConv_wxwin(wxFontEncoding enc)
{
m_enc = enc;
Init();
}
- size_t MB2WC(wchar_t *buf, const char *psz, size_t WXUNUSED(n))
+ size_t MB2WC(wchar_t *buf, const char *psz, size_t WXUNUSED(n)) const
{
size_t inbuf = strlen(psz);
if (buf)
return inbuf;
}
- size_t WC2MB(char *buf, const wchar_t *psz, size_t WXUNUSED(n))
+ size_t WC2MB(char *buf, const wchar_t *psz, size_t WXUNUSED(n)) const
{
const size_t inbuf = wxWcslen(psz);
if (buf)
return inbuf;
}
- bool usable() const { return m_ok; }
+ bool IsOk() const { return m_ok; }
public:
wxFontEncoding m_enc;
// were we initialized successfully?
bool m_ok;
- DECLARE_NO_COPY_CLASS(EC_CharSet)
+ DECLARE_NO_COPY_CLASS(wxMBConv_wxwin)
};
#endif // wxUSE_FONTMAP
-// ----------------------------------------------------------------------------
-// the function creating the wxCharacterSet for the specified charset on the
-// current system, trying all possibilities
-//
-// it uses the name if it is given or encoding if name == NULL
-// ----------------------------------------------------------------------------
-
-static wxCharacterSet *
-wxGetCharacterSet(const wxChar *name, wxFontEncoding encoding)
-{
- // check for the special case of ASCII charset
- if ( (!name && encoding == wxFONTENCODING_DEFAULT)
-#if wxUSE_FONTMAP
- || (name && wxFontMapper::Get()->
- CharsetToEncoding(name) == wxFONTENCODING_DEFAULT)
-#endif // wxUSE_FONTMAP
- )
- {
- // don't convert at all
- return NULL;
- }
-
- wxCharacterSet *cset = NULL;
-
- if (name)
- {
- if((wxStricmp(name, wxT("UTF8")) == 0) ||
- (wxStricmp(name, wxT("UTF-8")) == 0) ||
- encoding == wxFONTENCODING_UTF8 )
- {
- cset = new ID_CharSet(&wxConvUTF8);
- }
- else if((wxStricmp(name, wxT("UTF16")) == 0) ||
- (wxStricmp(name, wxT("UTF-16")) == 0) ||
- encoding == wxFONTENCODING_UTF16 )
- {
-#ifdef WORDS_BIGENDIAN
- cset = new ID_CharSet(&wxConvUTF16BE);
-#else
- cset = new ID_CharSet(&wxConvUTF16LE);
-#endif
- }
- else if((wxStricmp(name, wxT("UTF16BE")) == 0) ||
- (wxStricmp(name, wxT("UTF-16BE")) == 0) ||
- encoding == wxFONTENCODING_UTF16BE )
- {
- cset = new ID_CharSet(&wxConvUTF16BE);
- }
- else if((wxStricmp(name, wxT("UTF16LE")) == 0) ||
- (wxStricmp(name, wxT("UTF-16LE")) == 0) ||
- encoding == wxFONTENCODING_UTF16LE )
- {
- cset = new ID_CharSet(&wxConvUTF16LE);
- }
- else if((wxStricmp(name, wxT("UTF32")) == 0) ||
- (wxStricmp(name, wxT("UTF-32")) == 0) ||
- (wxStricmp(name, wxT("UCS4")) == 0) ||
- (wxStricmp(name, wxT("UCS-4")) == 0) ||
- encoding == wxFONTENCODING_UTF32 )
- {
-#ifdef WORDS_BIGENDIAN
- cset = new ID_CharSet(&wxConvUTF32BE);
-#else
- cset = new ID_CharSet(&wxConvUTF32LE);
-#endif
- }
- else if((wxStricmp(name, wxT("UTF32BE")) == 0) ||
- (wxStricmp(name, wxT("UTF-32BE")) == 0) ||
- (wxStricmp(name, wxT("UCS4BE")) == 0) ||
- (wxStricmp(name, wxT("UCS-4BE")) == 0) ||
- encoding == wxFONTENCODING_UTF32BE )
- {
- cset = new ID_CharSet(&wxConvUTF32BE);
- }
- else if((wxStricmp(name, wxT("UTF32LE")) == 0) ||
- (wxStricmp(name, wxT("UTF-32LE")) == 0) ||
- (wxStricmp(name, wxT("UCS4LE")) == 0) ||
- (wxStricmp(name, wxT("UCS-4LE")) == 0) ||
- encoding == wxFONTENCODING_UTF32 )
- {
- cset = new ID_CharSet(&wxConvUTF32LE);
- }
-#ifdef HAVE_ICONV
- else
- {
- cset = new IC_CharSet(name);
- }
-#endif // HAVE_ICONV
- }
-
- // 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__) && !defined(__WXUNIVERSAL__)
- cset = name ? new CP_CharSet(name) : new CP_CharSet(encoding);
- if ( cset->usable() )
- return cset;
-
- delete cset;
- cset = NULL;
-#endif // defined(__WIN32__) && !defined(__WXMICROWIN__) && !defined(__WXUNIVERSAL__)
-
-#if wxUSE_FONTMAP
- cset = name ? new EC_CharSet(name) : new EC_CharSet(encoding);
- if ( cset->usable() )
- return cset;
-
- delete cset;
- cset = NULL;
-#endif // wxUSE_FONTMAP
-
- wxLogError(_("Cannot convert from encoding '%s'!"),
- name ? name
- :
-#if wxUSE_FONTMAP
- wxFontMapper::GetEncodingDescription(encoding).c_str()
-#else // !wxUSE_FONTMAP
- wxString::Format(_T("%s"), encoding).c_str()
-#endif // wxUSE_FONTMAP/!wxUSE_FONTMAP
- );
-
- return NULL;
-}
-
// ============================================================================
// wxCSConv implementation
// ============================================================================
void wxCSConv::Init()
{
- m_name = (wxChar *)NULL;
- m_cset = (wxCharacterSet *) NULL;
- m_deferred = TRUE;
+ m_name = NULL;
+ m_convReal = NULL;
+ m_deferred = true;
}
wxCSConv::wxCSConv(const wxChar *charset)
{
Init();
- m_encoding = wxFONTENCODING_DEFAULT;
- SetName(charset);
+ if ( charset )
+ {
+ SetName(charset);
+ }
+
+ m_encoding = wxFONTENCODING_SYSTEM;
}
wxCSConv::wxCSConv(wxFontEncoding encoding)
{
+ if ( encoding == wxFONTENCODING_MAX || encoding == wxFONTENCODING_DEFAULT )
+ {
+ wxFAIL_MSG( _T("invalid encoding value in wxCSConv ctor") );
+
+ encoding = wxFONTENCODING_SYSTEM;
+ }
+
Init();
m_encoding = encoding;
void wxCSConv::Clear()
{
free(m_name);
- delete m_cset;
+ delete m_convReal;
m_name = NULL;
- m_cset = NULL;
+ m_convReal = NULL;
}
void wxCSConv::SetName(const wxChar *charset)
if (charset)
{
m_name = wxStrdup(charset);
- m_deferred = TRUE;
+ m_deferred = true;
}
}
-void wxCSConv::LoadNow()
+wxMBConv *wxCSConv::DoCreate() const
+{
+ // check for the special case of ASCII or ISO8859-1 charset: as we have
+ // special knowledge of it anyhow, we don't need to create a special
+ // conversion object
+ if ( m_encoding == wxFONTENCODING_ISO8859_1 )
+ {
+ // don't convert at all
+ return NULL;
+ }
+
+ // we trust OS to do conversion better than we can so try external
+ // conversion methods first
+ //
+ // the full order is:
+ // 1. OS conversion (iconv() under Unix or Win32 API)
+ // 2. hard coded conversions for UTF
+ // 3. wxEncodingConverter as fall back
+
+ // step (1)
+#ifdef HAVE_ICONV
+#if !wxUSE_FONTMAP
+ if ( m_name )
+#endif // !wxUSE_FONTMAP
+ {
+ wxString name(m_name);
+
+#if wxUSE_FONTMAP
+ if ( name.empty() )
+ name = wxFontMapper::Get()->GetEncodingName(m_encoding);
+#endif // wxUSE_FONTMAP
+
+ wxMBConv_iconv *conv = new wxMBConv_iconv(name);
+ if ( conv->IsOk() )
+ return conv;
+
+ delete conv;
+ }
+#endif // HAVE_ICONV
+
+#ifdef wxHAVE_WIN32_MB2WC
+ {
+ wxMBConv_win32 *conv = m_name ? new wxMBConv_win32(m_name)
+ : new wxMBConv_win32(m_encoding);
+ if ( conv->IsOk() )
+ return conv;
+
+ delete conv;
+ }
+#endif // wxHAVE_WIN32_MB2WC
+#if defined(__WXMAC__)
+ {
+ if ( m_name || ( m_encoding < wxFONTENCODING_UTF16BE ) )
+ {
+
+ wxMBConv_mac *conv = m_name ? new wxMBConv_mac(m_name)
+ : new wxMBConv_mac(m_encoding);
+ if ( conv->IsOk() )
+ return conv;
+
+ delete conv;
+ }
+ }
+#endif
+ // step (2)
+ wxFontEncoding enc = m_encoding;
+#if wxUSE_FONTMAP
+ if ( enc == wxFONTENCODING_SYSTEM && m_name )
+ {
+ // use "false" to suppress interactive dialogs -- we can be called from
+ // anywhere and popping up a dialog from here is the last thing we want to
+ // do
+ enc = wxFontMapper::Get()->CharsetToEncoding(m_name, false);
+ }
+#endif // wxUSE_FONTMAP
+
+ switch ( enc )
+ {
+ case wxFONTENCODING_UTF7:
+ return new wxMBConvUTF7;
+
+ case wxFONTENCODING_UTF8:
+ return new wxMBConvUTF8;
+
+ case wxFONTENCODING_UTF16BE:
+ return new wxMBConvUTF16BE;
+
+ case wxFONTENCODING_UTF16LE:
+ return new wxMBConvUTF16LE;
+
+ case wxFONTENCODING_UTF32BE:
+ return new wxMBConvUTF32BE;
+
+ case wxFONTENCODING_UTF32LE:
+ return new wxMBConvUTF32LE;
+
+ default:
+ // nothing to do but put here to suppress gcc warnings
+ ;
+ }
+
+ // step (3)
+#if wxUSE_FONTMAP
+ {
+ wxMBConv_wxwin *conv = m_name ? new wxMBConv_wxwin(m_name)
+ : new wxMBConv_wxwin(m_encoding);
+ if ( conv->IsOk() )
+ return conv;
+
+ delete conv;
+ }
+#endif // wxUSE_FONTMAP
+
+ // NB: This is a hack to prevent deadlock. What could otherwise happen
+ // in Unicode build: wxConvLocal creation ends up being here
+ // because of some failure and logs the error. But wxLog will try to
+ // attach timestamp, for which it will need wxConvLocal (to convert
+ // time to char* and then wchar_t*), but that fails, tries to log
+ // error, but wxLog has a (already locked) critical section that
+ // guards static buffer.
+ static bool alreadyLoggingError = false;
+ if (!alreadyLoggingError)
+ {
+ alreadyLoggingError = true;
+ wxLogError(_("Cannot convert from the charset '%s'!"),
+ m_name ? m_name
+ :
+#if wxUSE_FONTMAP
+ wxFontMapper::GetEncodingDescription(m_encoding).c_str()
+#else // !wxUSE_FONTMAP
+ wxString::Format(_("encoding %s"), m_encoding).c_str()
+#endif // wxUSE_FONTMAP/!wxUSE_FONTMAP
+ );
+ alreadyLoggingError = false;
+ }
+
+ return NULL;
+}
+
+void wxCSConv::CreateConvIfNeeded() const
{
if ( m_deferred )
{
- // it would probably be better to make GetSystemEncodingName() always
- // available (i.e. even when wxUSE_INTL == 0)?
+ wxCSConv *self = (wxCSConv *)this; // const_cast
+
#if wxUSE_INTL
- if ( !m_name && m_encoding == wxFONTENCODING_DEFAULT )
+ // if we don't have neither the name nor the encoding, use the default
+ // encoding for this system
+ if ( !m_name && m_encoding == wxFONTENCODING_SYSTEM )
{
- wxString name = wxLocale::GetSystemEncodingName();
- if ( !name.empty() )
- {
- SetName(name);
- }
+ self->m_name = wxStrdup(wxLocale::GetSystemEncodingName());
}
#endif // wxUSE_INTL
- // wxGetCharacterSet() complains about NULL name
- m_cset = wxGetCharacterSet(m_name, m_encoding);
- m_deferred = FALSE;
+ self->m_convReal = DoCreate();
+ self->m_deferred = false;
}
}
size_t wxCSConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
- ((wxCSConv *)this)->LoadNow(); // discard constness
+ CreateConvIfNeeded();
- if (m_cset)
- return m_cset->MB2WC(buf, psz, n);
+ if (m_convReal)
+ return m_convReal->MB2WC(buf, psz, n);
// latin-1 (direct)
size_t len = strlen(psz);
size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const
{
- ((wxCSConv *)this)->LoadNow(); // discard constness
+ CreateConvIfNeeded();
- if (m_cset)
- return m_cset->WC2MB(buf, psz, n);
+ if (m_convReal)
+ return m_convReal->WC2MB(buf, psz, n);
// latin-1 (direct)
const size_t len = wxWcslen(psz);
if (buf)
{
for (size_t c = 0; c <= len; c++)
- buf[c] = (psz[c] > 0xff) ? '?' : psz[c];
+ {
+ if (psz[c] > 0xFF)
+ return (size_t)-1;
+ buf[c] = psz[c];
+ }
+ }
+ else
+ {
+ for (size_t c = 0; c <= len; c++)
+ {
+ if (psz[c] > 0xFF)
+ return (size_t)-1;
+ }
}
return len;
}
-#endif // wxUSE_WCHAR_T
+// ----------------------------------------------------------------------------
+// globals
+// ----------------------------------------------------------------------------
+
+#ifdef __WINDOWS__
+ static wxMBConv_win32 wxConvLibcObj;
+#else
+ static wxMBConvLibc wxConvLibcObj;
+#endif
+
+static wxCSConv wxConvLocalObj(wxFONTENCODING_SYSTEM);
+static wxCSConv wxConvISO8859_1Obj(wxFONTENCODING_ISO8859_1);
+static wxMBConvUTF7 wxConvUTF7Obj;
+static wxMBConvUTF8 wxConvUTF8Obj;
+
+
+WXDLLIMPEXP_DATA_BASE(wxMBConv&) wxConvLibc = wxConvLibcObj;
+WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvLocal = wxConvLocalObj;
+WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvISO8859_1 = wxConvISO8859_1Obj;
+WXDLLIMPEXP_DATA_BASE(wxMBConvUTF7&) wxConvUTF7 = wxConvUTF7Obj;
+WXDLLIMPEXP_DATA_BASE(wxMBConvUTF8&) wxConvUTF8 = wxConvUTF8Obj;
+WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent = &wxConvLibcObj;
+
+#else // !wxUSE_WCHAR_T
+
+// stand-ins in absence of wchar_t
+WXDLLIMPEXP_DATA_BASE(wxMBConv) wxConvLibc,
+ wxConvISO8859_1,
+ wxConvLocal,
+ wxConvUTF8;
+
+#endif // wxUSE_WCHAR_T/!wxUSE_WCHAR_T
projects / wxWidgets.git / blobdiff