// Modified by:
// Created: 29/01/98
// RCS-ID: $Id$
-// Copyright: (c) 1999 Ove Kaaven, Robert Roebling, Vadim Zeitlin, Vaclav Slavik
-// Licence: wxWindows license
+// Copyright: (c) 1999 Ove Kaaven, Robert Roebling, Vaclav Slavik
+// (c) 2000-2003 Vadim Zeitlin
+// Licence: wxWindows licence
/////////////////////////////////////////////////////////////////////////////
// ============================================================================
// headers
// ----------------------------------------------------------------------------
-#ifdef __GNUG__
+#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA)
#pragma implementation "strconv.h"
#endif
#include "wx/log.h"
#endif // WX_PRECOMP
+#include "wx/strconv.h"
+
+#if wxUSE_WCHAR_T
+
#ifdef __WXMSW__
#include "wx/msw/private.h"
#endif
+#ifndef __WXWINCE__
#include <errno.h>
+#endif
+
#include <ctype.h>
#include <string.h>
#include <stdlib.h>
-#include "wx/module.h"
-#include "wx/strconv.h"
-
-// ----------------------------------------------------------------------------
-// globals
-// ----------------------------------------------------------------------------
-
-#if wxUSE_WCHAR_T
- WXDLLEXPORT_DATA(wxMBConv) wxConvLibc;
- WXDLLEXPORT_DATA(wxCSConv) wxConvLocal((const wxChar *)NULL);
- WXDLLEXPORT_DATA(wxCSConv) wxConvISO8859_1(_T("iso-8859-1"));
-#else
- // stand-ins in absence of wchar_t
- WXDLLEXPORT_DATA(wxMBConv) wxConvLibc,
- wxConvFile,
- wxConvISO8859_1,
- wxConvLocal;
-#endif // wxUSE_WCHAR_T
-
-WXDLLEXPORT_DATA(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
// ============================================================================
// ============================================================================
// ----------------------------------------------------------------------------
-// UTF-16 en/decoding
+// UTF-16 en/decoding to/from UCS-4
// ----------------------------------------------------------------------------
-#ifdef WC_UTF16
-static size_t encode_utf16(wxUint32 input, wchar_t *output)
+static size_t encode_utf16(wxUint32 input, wxUint16 *output)
{
if (input<=0xffff)
{
- if (output) *output++ = (wchar_t) input;
+ if (output)
+ *output = (wxUint16) input;
return 1;
}
else if (input>=0x110000)
{
if (output)
{
- *output++ = (wchar_t) ((input >> 10)+0xd7c0);
- *output++ = (wchar_t) ((input&0x3ff)+0xdc00);
+ *output++ = (wxUint16) ((input >> 10)+0xd7c0);
+ *output = (wxUint16) ((input&0x3ff)+0xdc00);
}
return 2;
}
}
-static size_t decode_utf16(const wchar_t* input, wxUint32& output)
+static size_t decode_utf16(const wxUint16* input, wxUint32& output)
{
if ((*input<0xd800) || (*input>0xdfff))
{
}
}
-#endif // WC_UTF16
// ----------------------------------------------------------------------------
// 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 )
size_t nLen = WC2MB(NULL, pwz, 0);
if ( nLen != (size_t)-1 )
{
- wxCharBuffer buf(nLen);
- WC2MB(buf.data(), pwz, nLen + 1);
+ wxCharBuffer buf(nLen+3); // space for a wxUint32 trailing zero
+ WC2MB(buf.data(), pwz, nLen + 4);
return buf;
}
}
// ----------------------------------------------------------------------------
-// standard gdk conversion
+// wxMBConvLibc
// ----------------------------------------------------------------------------
-#ifdef __WXGTK12__
-
-WXDLLEXPORT_DATA(wxMBConvGdk) wxConvGdk;
-
-#include <gdk/gdk.h>
-
-size_t wxMBConvGdk::MB2WC(wchar_t *buf, const char *psz, size_t n) const
+size_t wxMBConvLibc::MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
- if (buf)
- {
- return gdk_mbstowcs((GdkWChar *)buf, psz, n);
- }
- else
- {
- GdkWChar *nbuf = new GdkWChar[n=strlen(psz)];
- size_t len = gdk_mbstowcs(nbuf, psz, n);
- delete[] nbuf;
- return len;
- }
+ return wxMB2WC(buf, psz, n);
}
-size_t wxMBConvGdk::WC2MB(char *buf, const wchar_t *psz, size_t n) const
+size_t wxMBConvLibc::WC2MB(char *buf, const wchar_t *psz, size_t n) const
{
- char *mbstr = gdk_wcstombs((GdkWChar *)psz);
- size_t len = mbstr ? strlen(mbstr) : 0;
- if (buf)
- {
- if (len > n)
- len = n;
- memcpy(buf, psz, len);
- if (len < n)
- buf[len] = 0;
- }
- return len;
+ return wxWC2MB(buf, psz, n);
}
-#endif // GTK > 1.0
-
// ----------------------------------------------------------------------------
// UTF-7
// ----------------------------------------------------------------------------
-WXDLLEXPORT_DATA(wxMBConvUTF7) wxConvUTF7;
-
#if 0
static char utf7_setD[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ"
"abcdefghijklmnopqrstuvwxyz"
// UTF-8
// ----------------------------------------------------------------------------
-WXDLLEXPORT_DATA(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;
}
if (buf && (len<n)) *buf = 0;
-
+
return len;
}
-// ============================================================================
-// wxCharacterSet and derived classes
-// ============================================================================
+
+
// ----------------------------------------------------------------------------
-// wxCharacterSet is the ABC for the classes below
+// UTF-16
// ----------------------------------------------------------------------------
-class wxCharacterSet
+#ifdef WORDS_BIGENDIAN
+ #define wxMBConvUTF16straight wxMBConvUTF16BE
+ #define wxMBConvUTF16swap wxMBConvUTF16LE
+#else
+ #define wxMBConvUTF16swap wxMBConvUTF16BE
+ #define wxMBConvUTF16straight wxMBConvUTF16LE
+#endif
+
+
+#ifdef WC_UTF16
+
+// copy 16bit MB to 16bit String
+size_t wxMBConvUTF16straight::MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
-public:
- 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;
-};
+ size_t len=0;
+
+ while (*(wxUint16*)psz && (!buf || len < n))
+ {
+ if (buf)
+ *buf++ = *(wxUint16*)psz;
+ len++;
+
+ psz += sizeof(wxUint16);
+ }
+ if (buf && len<n) *buf=0;
+
+ return len;
+}
+
+
+// copy 16bit String to 16bit MB
+size_t wxMBConvUTF16straight::WC2MB(char *buf, const wchar_t *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*psz && (!buf || len < n))
+ {
+ if (buf)
+ {
+ *(wxUint16*)buf = *psz;
+ buf += sizeof(wxUint16);
+ }
+ len += sizeof(wxUint16);
+ psz++;
+ }
+ if (buf && len<=n-sizeof(wxUint16)) *(wxUint16*)buf=0;
+
+ return len;
+}
+
+
+// swap 16bit MB to 16bit String
+size_t wxMBConvUTF16swap::MB2WC(wchar_t *buf, const char *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*(wxUint16*)psz && (!buf || len < n))
+ {
+ if (buf)
+ {
+ ((char *)buf)[0] = psz[1];
+ ((char *)buf)[1] = psz[0];
+ buf++;
+ }
+ len++;
+ psz += sizeof(wxUint16);
+ }
+ if (buf && len<n) *buf=0;
+
+ return len;
+}
+
+
+// swap 16bit MB to 16bit String
+size_t wxMBConvUTF16swap::WC2MB(char *buf, const wchar_t *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*psz && (!buf || len < n))
+ {
+ if (buf)
+ {
+ *buf++ = ((char*)psz)[1];
+ *buf++ = ((char*)psz)[0];
+ }
+ len += sizeof(wxUint16);
+ psz++;
+ }
+ if (buf && len<=n-sizeof(wxUint16)) *(wxUint16*)buf=0;
+
+ return len;
+}
+
+
+#else // WC_UTF16
+
+
+// copy 16bit MB to 32bit String
+size_t wxMBConvUTF16straight::MB2WC(wchar_t *buf, const char *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*(wxUint16*)psz && (!buf || len < n))
+ {
+ wxUint32 cc;
+ size_t pa=decode_utf16((wxUint16*)psz, cc);
+ if (pa == (size_t)-1)
+ return pa;
+
+ if (buf)
+ *buf++ = cc;
+ len++;
+ psz += pa * sizeof(wxUint16);
+ }
+ if (buf && len<n) *buf=0;
+
+ return len;
+}
+
+
+// copy 32bit String to 16bit MB
+size_t wxMBConvUTF16straight::WC2MB(char *buf, const wchar_t *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*psz && (!buf || len < n))
+ {
+ wxUint16 cc[2];
+ size_t pa=encode_utf16(*psz, cc);
+
+ if (pa == (size_t)-1)
+ return pa;
+
+ if (buf)
+ {
+ *(wxUint16*)buf = cc[0];
+ buf += sizeof(wxUint16);
+ if (pa > 1)
+ {
+ *(wxUint16*)buf = cc[1];
+ buf += sizeof(wxUint16);
+ }
+ }
+
+ len += pa*sizeof(wxUint16);
+ psz++;
+ }
+ if (buf && len<=n-sizeof(wxUint16)) *(wxUint16*)buf=0;
+
+ return len;
+}
+
+
+// swap 16bit MB to 32bit String
+size_t wxMBConvUTF16swap::MB2WC(wchar_t *buf, const char *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*(wxUint16*)psz && (!buf || len < n))
+ {
+ wxUint32 cc;
+ char tmp[4];
+ tmp[0]=psz[1]; tmp[1]=psz[0];
+ tmp[2]=psz[3]; tmp[3]=psz[2];
+
+ size_t pa=decode_utf16((wxUint16*)tmp, cc);
+ if (pa == (size_t)-1)
+ return pa;
+
+ if (buf)
+ *buf++ = cc;
+
+ len++;
+ psz += pa * sizeof(wxUint16);
+ }
+ if (buf && len<n) *buf=0;
+
+ return len;
+}
+
+
+// swap 32bit String to 16bit MB
+size_t wxMBConvUTF16swap::WC2MB(char *buf, const wchar_t *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*psz && (!buf || len < n))
+ {
+ wxUint16 cc[2];
+ size_t pa=encode_utf16(*psz, cc);
+
+ if (pa == (size_t)-1)
+ return pa;
+
+ if (buf)
+ {
+ *buf++ = ((char*)cc)[1];
+ *buf++ = ((char*)cc)[0];
+ if (pa > 1)
+ {
+ *buf++ = ((char*)cc)[3];
+ *buf++ = ((char*)cc)[2];
+ }
+ }
+
+ len += pa*sizeof(wxUint16);
+ psz++;
+ }
+ if (buf && len<=n-sizeof(wxUint16)) *(wxUint16*)buf=0;
+
+ return len;
+}
+
+#endif // WC_UTF16
+
// ----------------------------------------------------------------------------
-// ID_CharSet: implementation of wxCharacterSet using an existing wxMBConv
+// UTF-32
// ----------------------------------------------------------------------------
-class ID_CharSet : public wxCharacterSet
+#ifdef WORDS_BIGENDIAN
+#define wxMBConvUTF32straight wxMBConvUTF32BE
+#define wxMBConvUTF32swap wxMBConvUTF32LE
+#else
+#define wxMBConvUTF32swap wxMBConvUTF32BE
+#define wxMBConvUTF32straight wxMBConvUTF32LE
+#endif
+
+
+WXDLLIMPEXP_DATA_BASE(wxMBConvUTF32LE) wxConvUTF32LE;
+WXDLLIMPEXP_DATA_BASE(wxMBConvUTF32BE) wxConvUTF32BE;
+
+
+#ifdef WC_UTF16
+
+// copy 32bit MB to 16bit String
+size_t wxMBConvUTF32straight::MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
-public:
- ID_CharSet(const wxChar *name, wxMBConv *cnv)
- : wxCharacterSet(name), work(cnv) {}
+ size_t len=0;
- size_t MB2WC(wchar_t *buf, const char *psz, size_t n)
- { return work ? work->MB2WC(buf,psz,n) : (size_t)-1; }
+ while (*(wxUint32*)psz && (!buf || len < n))
+ {
+ wxUint16 cc[2];
- size_t WC2MB(char *buf, const wchar_t *psz, size_t n)
- { return work ? work->WC2MB(buf,psz,n) : (size_t)-1; }
+ size_t pa=encode_utf16(*(wxUint32*)psz, cc);
+ if (pa == (size_t)-1)
+ return pa;
- bool usable() const
- { return work!=NULL; }
-public:
- wxMBConv*work;
-};
+ if (buf)
+ {
+ *buf++ = cc[0];
+ if (pa > 1)
+ *buf++ = cc[1];
+ }
+ len += pa;
+ psz += sizeof(wxUint32);
+ }
+ if (buf && len<n) *buf=0;
+
+ return len;
+}
+
+
+// copy 16bit String to 32bit MB
+size_t wxMBConvUTF32straight::WC2MB(char *buf, const wchar_t *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*psz && (!buf || len < n))
+ {
+ wxUint32 cc;
+
+ // cast is ok for WC_UTF16
+ size_t pa = decode_utf16((const wxUint16 *)psz, cc);
+ if (pa == (size_t)-1)
+ return pa;
+
+ if (buf)
+ {
+ *(wxUint32*)buf = cc;
+ buf += sizeof(wxUint32);
+ }
+ len += sizeof(wxUint32);
+ psz += pa;
+ }
+
+ if (buf && len<=n-sizeof(wxUint32))
+ *(wxUint32*)buf=0;
+
+ return len;
+}
+
+
+
+// swap 32bit MB to 16bit String
+size_t wxMBConvUTF32swap::MB2WC(wchar_t *buf, const char *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*(wxUint32*)psz && (!buf || len < n))
+ {
+ char tmp[4];
+ tmp[0] = psz[3]; tmp[1] = psz[2];
+ tmp[2] = psz[1]; tmp[3] = psz[0];
+
+
+ wxUint16 cc[2];
+
+ size_t pa=encode_utf16(*(wxUint32*)tmp, cc);
+ if (pa == (size_t)-1)
+ return pa;
+
+ if (buf)
+ {
+ *buf++ = cc[0];
+ if (pa > 1)
+ *buf++ = cc[1];
+ }
+ len += pa;
+ psz += sizeof(wxUint32);
+ }
+
+ if (buf && len<n)
+ *buf=0;
+
+ return len;
+}
+
+
+// swap 16bit String to 32bit MB
+size_t wxMBConvUTF32swap::WC2MB(char *buf, const wchar_t *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*psz && (!buf || len < n))
+ {
+ char cc[4];
+
+ // cast is ok for WC_UTF16
+ size_t pa=decode_utf16((const wxUint16 *)psz, *(wxUint32*)cc);
+ if (pa == (size_t)-1)
+ return pa;
+
+ if (buf)
+ {
+ *buf++ = cc[3];
+ *buf++ = cc[2];
+ *buf++ = cc[1];
+ *buf++ = cc[0];
+ }
+ len += sizeof(wxUint32);
+ psz += pa;
+ }
+
+ if (buf && len<=n-sizeof(wxUint32))
+ *(wxUint32*)buf=0;
+
+ return len;
+}
+
+#else // WC_UTF16
+
+
+// copy 32bit MB to 32bit String
+size_t wxMBConvUTF32straight::MB2WC(wchar_t *buf, const char *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*(wxUint32*)psz && (!buf || len < n))
+ {
+ if (buf)
+ *buf++ = *(wxUint32*)psz;
+ len++;
+ psz += sizeof(wxUint32);
+ }
+
+ if (buf && len<n)
+ *buf=0;
+
+ return len;
+}
+
+
+// copy 32bit String to 32bit MB
+size_t wxMBConvUTF32straight::WC2MB(char *buf, const wchar_t *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*psz && (!buf || len < n))
+ {
+ if (buf)
+ {
+ *(wxUint32*)buf = *psz;
+ buf += sizeof(wxUint32);
+ }
+
+ len += sizeof(wxUint32);
+ psz++;
+ }
+
+ if (buf && len<=n-sizeof(wxUint32))
+ *(wxUint32*)buf=0;
+
+ return len;
+}
+
+
+// swap 32bit MB to 32bit String
+size_t wxMBConvUTF32swap::MB2WC(wchar_t *buf, const char *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*(wxUint32*)psz && (!buf || len < n))
+ {
+ if (buf)
+ {
+ ((char *)buf)[0] = psz[3];
+ ((char *)buf)[1] = psz[2];
+ ((char *)buf)[2] = psz[1];
+ ((char *)buf)[3] = psz[0];
+ buf++;
+ }
+ len++;
+ psz += sizeof(wxUint32);
+ }
+
+ if (buf && len<n)
+ *buf=0;
+
+ return len;
+}
+
+
+// swap 32bit String to 32bit MB
+size_t wxMBConvUTF32swap::WC2MB(char *buf, const wchar_t *psz, size_t n) const
+{
+ size_t len=0;
+
+ while (*psz && (!buf || len < n))
+ {
+ if (buf)
+ {
+ *buf++ = ((char *)psz)[3];
+ *buf++ = ((char *)psz)[2];
+ *buf++ = ((char *)psz)[1];
+ *buf++ = ((char *)psz)[0];
+ }
+ len += sizeof(wxUint32);
+ psz++;
+ }
+
+ if (buf && len<=n-sizeof(wxUint32))
+ *(wxUint32*)buf=0;
+
+ return len;
+}
+
+
+#endif // WC_UTF16
// ============================================================================
#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)
- : wxCharacterSet(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;
// convert to native endianness
WC_BSWAP(buf /* _not_ bufPtr */, res)
}
-
+
// NB: iconv was given only strlen(psz) characters on input, and so
// it couldn't convert the trailing zero. Let's do it ourselves
// if there's some room left for it in the output buffer.
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;
cres = iconv( w2m, ICONV_CHAR_CAST(&psz), &inbuf, &buf, &outbuf );
res = n-outbuf;
-
+
// NB: iconv was given only wcslen(psz) characters on input, and so
// it couldn't convert the trailing zero. Let's do it ourselves
// if there's some room left for it in the output buffer.
#endif // HAVE_ICONV
+
// ============================================================================
// Win32 conversion classes
// ============================================================================
-#if defined(__WIN32__) && !defined(__WXMICROWIN__)
-
-#ifdef __WXWINE__
- #define WINE_CAST (WCHAR *)
-#else
- #define WINE_CAST
-#endif
+#ifdef wxHAVE_WIN32_MB2WC
-extern long wxCharsetToCodepage(const wxChar *charset); // from utils.cpp
+// 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)
- : wxCharacterSet(name)
- {
- m_CodePage = wxCharsetToCodepage(name);
- }
+ wxMBConv_win32()
+ {
+ m_CodePage = CP_ACP;
+ }
+
+ wxMBConv_win32(const wxChar* name)
+ {
+ m_CodePage = wxCharsetToCodepage(name);
+ }
+
+ wxMBConv_win32(wxFontEncoding encoding)
+ {
+ m_CodePage = wxEncodingToCodepage(encoding);
+ }
- size_t MB2WC(wchar_t *buf, const char *psz, size_t n)
+ size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
const size_t len = ::MultiByteToWideChar
(
0, // flags (none)
psz, // input string
-1, // its length (NUL-terminated)
- WINE_CAST buf, // output string
+ buf, // output string
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
(
m_CodePage, // code page
0, // flags (none)
- WINE_CAST psz, // input string
+ psz, // input string
-1, // it is (wide) NUL-terminated
buf, // output buffer
buf ? n : 0, // and its size
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 // __WIN32__
+
+#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()
+ {
+ m_ok = m2w.Init(m_enc, wxFONTENCODING_UNICODE) &&
+ w2m.Init(wxFONTENCODING_UNICODE, m_enc);
+ }
+
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)
+ wxMBConv_wxwin(const wxChar* name)
{
if (name)
- enc = wxFontMapper::Get()->CharsetToEncoding(name, FALSE);
+ m_enc = wxFontMapper::Get()->CharsetToEncoding(name, false);
+ else
+ m_enc = wxFONTENCODING_SYSTEM;
- m_ok = m2w.Init(enc, wxFONTENCODING_UNICODE) &&
- w2m.Init(wxFONTENCODING_UNICODE, enc);
+ Init();
}
- size_t MB2WC(wchar_t *buf, const char *psz, size_t WXUNUSED(n))
+ wxMBConv_wxwin(wxFontEncoding enc)
+ {
+ m_enc = enc;
+
+ Init();
+ }
+
+ 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 enc;
+ wxFontEncoding m_enc;
wxEncodingConverter m2w, w2m;
// were we initialized successfully?
bool m_ok;
+
+ 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
-// ----------------------------------------------------------------------------
+// ============================================================================
+// wxCSConv implementation
+// ============================================================================
-static wxCharacterSet *wxGetCharacterSet(const wxChar *name)
+void wxCSConv::Init()
{
- // check for the special case of ASCII charset
-#if wxUSE_FONTMAP
- if ( wxFontMapper::Get()->CharsetToEncoding(name) == wxFONTENCODING_DEFAULT )
-#else // wxUSE_FONTMAP
- if ( !name )
-#endif // wxUSE_FONTMAP/!wxUSE_FONTMAP
- {
- // don't convert at all
- return NULL;
- }
-
- // the test above must have taken care of this case
- wxCHECK_MSG( name, NULL, _T("NULL name must be wxFONTENCODING_DEFAULT") );
+ m_name = NULL;
+ m_convReal = NULL;
+ m_deferred = true;
+}
- wxCharacterSet *cset;
+wxCSConv::wxCSConv(const wxChar *charset)
+{
+ Init();
- if ( wxStricmp(name, wxT("UTF8")) == 0 || wxStricmp(name, wxT("UTF-8")) == 0)
- {
- cset = new ID_CharSet(name, &wxConvUTF8);
- }
- else
+ if ( charset )
{
-#ifdef HAVE_ICONV
- cset = new IC_CharSet(name);
-#else // !HAVE_ICONV
- cset = NULL;
-#endif // HAVE_ICONV/!HAVE_ICONV
+ SetName(charset);
}
- // it can only be NULL in this case
-#ifndef HAVE_ICONV
- if ( cset )
-#endif // !HAVE_ICONV
+ m_encoding = wxFONTENCODING_SYSTEM;
+}
+
+wxCSConv::wxCSConv(wxFontEncoding encoding)
+{
+ if ( encoding == wxFONTENCODING_MAX || encoding == wxFONTENCODING_DEFAULT )
{
- if ( cset->usable() )
- return cset;
+ wxFAIL_MSG( _T("invalid encoding value in wxCSConv ctor") );
- delete cset;
- cset = NULL;
+ encoding = wxFONTENCODING_SYSTEM;
}
-#if defined(__WIN32__) && !defined(__WXMICROWIN__)
- cset = new CP_CharSet(name);
- if ( cset->usable() )
- return cset;
-
- delete cset;
- cset = NULL;
-#endif // __WIN32__
-
-#if wxUSE_FONTMAP
- cset = new EC_CharSet(name);
- if ( cset->usable() )
- return cset;
+ Init();
- delete cset;
- cset = NULL;
-#endif // wxUSE_FONTMAP
-
- wxLogError(_("Cannot convert from encoding '%s'!"), name);
-
- return NULL;
-}
-
-// ============================================================================
-// wxCSConv implementation
-// ============================================================================
-
-wxCSConv::wxCSConv(const wxChar *charset)
-{
- m_name = (wxChar *)NULL;
- m_cset = (wxCharacterSet *) NULL;
- m_deferred = TRUE;
-
- SetName(charset);
+ m_encoding = encoding;
}
wxCSConv::~wxCSConv()
}
wxCSConv::wxCSConv(const wxCSConv& conv)
- : wxMBConv()
+ : wxMBConv()
{
- Clear();
+ Init();
+
SetName(conv.m_name);
+ m_encoding = conv.m_encoding;
}
wxCSConv& wxCSConv::operator=(const wxCSConv& conv)
{
Clear();
+
SetName(conv.m_name);
+ m_encoding = conv.m_encoding;
+
return *this;
}
void wxCSConv::Clear()
{
- if (m_name)
- free(m_name);
- if (m_cset)
- delete m_cset;
+ free(m_name);
+ 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;
+ }
+}
+
+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::LoadNow()
+void wxCSConv::CreateConvIfNeeded() const
{
- if (m_deferred)
+ if ( m_deferred )
{
- if ( !m_name )
+ wxCSConv *self = (wxCSConv *)this; // const_cast
+
+#if wxUSE_INTL
+ // 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 = m_name ? wxGetCharacterSet(m_name) : NULL;
- 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