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

/////////////////////////////////////////////////////////////////////////////
// Name:        strconv.cpp
// Purpose:     Unicode conversion classes
// Author:      Ove Kaaven, Robert Roebling, Vadim Zeitlin
// Modified by:
// Created:     29/01/98
// RCS-ID:      $Id$
// Copyright:   (c) 1999 Ove Kaaven, Robert Roebling, Vadim Zeitlin
// Licence:     wxWindows license
/////////////////////////////////////////////////////////////////////////////

// ============================================================================
// declarations
// ============================================================================

// ----------------------------------------------------------------------------
// headers
// ----------------------------------------------------------------------------

#ifdef __GNUG__
  #pragma implementation "strconv.h"
#endif

// For compilers that support precompilation, includes "wx.h".
#include "wx/wxprec.h"

#ifdef __BORLANDC__
  #pragma hdrstop
#endif

#ifdef __WXMSW__
  #include "wx/msw/private.h"
#endif

#include <errno.h>
#include <ctype.h>
#include <string.h>
#include <stdlib.h>

#ifdef __SALFORDC__
  #include <clib.h>
#endif

#ifdef HAVE_ICONV_H
  #include <iconv.h>
#endif
#ifdef HAVE_LANGINFO_H
  #include <langinfo.h>
#endif

#ifdef __WXMSW__
  #include <windows.h>
#endif

#include "wx/debug.h"
#include "wx/strconv.h"

#if defined(WORDS_BIGENDIAN) || defined(__STDC_ISO_10646__)
#define BSWAP_UCS4(str, len)
#define BSWAP_UCS2(str, len)
#else
#define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT32_SWAP_ALWAYS(str[_c]); }
#define BSWAP_UCS2(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT16_SWAP_ALWAYS(str[_c]); }
#define WC_NEED_BSWAP
#endif
#define BSWAP_UTF32(str, len) BSWAP_UCS4(str, len)
#define BSWAP_UTF16(str, len) BSWAP_UCS2(str, len)

#if SIZEOF_WCHAR_T == 4
#define WC_NAME "UCS4"
#define WC_BSWAP BSWAP_UCS4
#elif SIZEOF_WCHAR_T == 2
#define WC_NAME "UTF16"
#define WC_BSWAP BSWAP_UTF16
#define WC_UTF16
#endif

// ----------------------------------------------------------------------------
// globals
// ----------------------------------------------------------------------------

WXDLLEXPORT_DATA(wxMBConv *) wxConvCurrent = &wxConvLibc;

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

#if wxUSE_WCHAR_T

#ifdef WC_UTF16

static size_t encode_utf16(wxUint32 input,wxUint16*output)
{
  if (input<=0xffff) {
    if (output) *output++ = input;
    return 1;
  } else
  if (input>=0x110000) {
    return (size_t)-1;
  } else {
    if (output) {
      *output++ = (input >> 10)+0xd7c0;
      *output++ = (input&0x3ff)+0xdc00;
    }
    return 2;
  }
}

static size_t decode_utf16(wxUint16*input,wxUint32&output)
{
  if ((*input<0xd800) || (*input>0xdfff)) {
    output = *input;
    return 1;
  } else
  if ((input[1]<0xdc00) || (input[1]>=0xdfff)) {
    output = *input;
    return (size_t)-1;
  } else {
    output = ((input[0] - 0xd7c0) << 10) + (input[1] - 0xdc00);
    return 2;
  }
}

#endif // WC_UTF16

// ----------------------------------------------------------------------------
// 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::WC2MB(char *buf, const wchar_t *psz, size_t n) const
{
    return wxWC2MB(buf, psz, n);
}

const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const
{
    if (psz)
    {
        size_t nLen = MB2WC((wchar_t *) NULL, psz, 0);
        if (nLen == (size_t)-1)
            return wxWCharBuffer((wchar_t *) NULL);
        wxWCharBuffer buf(nLen);
        MB2WC((wchar_t *)(const wchar_t *) buf, psz, nLen);
        return buf;
    }
    else
        return wxWCharBuffer((wchar_t *) NULL);
}

const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *psz) const
{
    if (psz)
    {
        size_t nLen = WC2MB((char *) NULL, psz, 0);
        if (nLen == (size_t)-1)
            return wxCharBuffer((char *) NULL);
        wxCharBuffer buf(nLen);
        WC2MB((char *)(const char *) buf, psz, nLen);
        return buf;
    }
    else
        return wxCharBuffer((char *) NULL);
}

// ----------------------------------------------------------------------------
// standard file conversion
// ----------------------------------------------------------------------------

WXDLLEXPORT_DATA(wxMBConvFile) wxConvFile;

// just use the libc conversion for now
size_t wxMBConvFile::MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
    return wxMB2WC(buf, psz, n);
}

size_t wxMBConvFile::WC2MB(char *buf, const wchar_t *psz, size_t n) const
{
    return wxWC2MB(buf, psz, n);
}

// ----------------------------------------------------------------------------
// standard gdk conversion
// ----------------------------------------------------------------------------

#ifdef __WXGTK12__

WXDLLEXPORT_DATA(wxMBConvGdk) wxConvGdk;

#include <gdk/gdk.h>

size_t wxMBConvGdk::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;
  }
}

size_t wxMBConvGdk::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;
}

#endif // GTK > 1.0

// ----------------------------------------------------------------------------
// UTF-7
// ----------------------------------------------------------------------------

WXDLLEXPORT_DATA(wxMBConvUTF7) wxConvUTF7;

#if 0
static char utf7_setD[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ"
                        "abcdefghijklmnopqrstuvwxyz"
                        "0123456789'(),-./:?";
static char utf7_setO[]="!\"#$%&*;<=>@[]^_`{|}";
static char utf7_setB[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ"
                        "abcdefghijklmnopqrstuvwxyz"
                        "0123456789+/";
#endif

// TODO: write actual implementations of UTF-7 here
size_t wxMBConvUTF7::MB2WC(wchar_t * WXUNUSED(buf),
                           const char * WXUNUSED(psz),
                           size_t WXUNUSED(n)) const
{
  return 0;
}

size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf),
                           const wchar_t * WXUNUSED(psz),
                           size_t WXUNUSED(n)) const
{
  return 0;
}

// ----------------------------------------------------------------------------
// UTF-8
// ----------------------------------------------------------------------------

WXDLLEXPORT_DATA(wxMBConvUTF8) wxConvUTF8;

static wxUint32 utf8_max[]={0x7f,0x7ff,0xffff,0x1fffff,0x3ffffff,0x7fffffff,0xffffffff};

size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
  size_t len = 0;

  while (*psz && ((!buf) || (len<n))) {
    unsigned char cc=*psz++, fc=cc;
    unsigned cnt;
    for (cnt=0; fc&0x80; cnt++) fc<<=1;
    if (!cnt) {
      // plain ASCII char
      if (buf) *buf++=cc;
      len++;
    } else {
      cnt--;
      if (!cnt) {
        // invalid UTF-8 sequence
        return (size_t)-1;
      } else {
        unsigned ocnt=cnt-1;
        wxUint32 res=cc&(0x3f>>cnt);
        while (cnt--) {
          cc = *psz++;
          if ((cc&0xC0)!=0x80) {
            // invalid UTF-8 sequence
            return (size_t)-1;
          }
          res=(res<<6)|(cc&0x3f);
        }
        if (res<=utf8_max[ocnt]) {
          // illegal UTF-8 encoding
          return (size_t)-1;
        }
#ifdef WC_UTF16
        size_t pa = encode_utf16(res, buf);
        if (pa == (size_t)-1)
          return (size_t)-1;
        if (buf) buf+=pa;
        len+=pa;
#else
        if (buf) *buf++=res;
        len++;
#endif
      }
    }
  }
  if (buf && (len<n)) *buf = 0;
  return len;
}

size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const
{
  size_t len = 0;

  while (*psz && ((!buf) || (len<n))) {
    wxUint32 cc;
#ifdef WC_UTF16
    size_t pa = decode_utf16(psz,cc);
    psz += (pa == (size_t)-1) ? 1 : pa;
#else
    cc=(*psz++)&0x7fffffff;
#endif
    unsigned cnt;
    for (cnt=0; cc>utf8_max[cnt]; cnt++);
    if (!cnt) {
      // plain ASCII char
      if (buf) *buf++=cc;
      len++;
    } else {
      len+=cnt+1;
      if (buf) {
        *buf++=(-128>>cnt)|((cc>>(cnt*6))&(0x3f>>cnt));
        while (cnt--)
          *buf++=0x80|((cc>>(cnt*6))&0x3f);
      }
    }
  }
  if (buf && (len<n)) *buf = 0;
  return len;
}

// ----------------------------------------------------------------------------
// specified character set
// ----------------------------------------------------------------------------

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

#ifdef __WIN32__
#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);

    /* 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.QueryValue(wxT("InternetEncoding"),&CP)) break;
    // no encoding, see if it's an alias
    if (!key.QueryValue(wxT("AliasForCharset"),cn)) break;
  } while (1);
  return CP;
}
#endif

class wxCharacterSet
{
public:
  const wxChar*cname;
  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; }
};

class ID_CharSet : public wxCharacterSet
{
public:
  wxMBConv*work;
  ID_CharSet(const wxChar*name,wxMBConv*cnv) : wxCharacterSet(name), 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() { return work!=NULL; }
};

#ifdef HAVE_ICONV_H
class IC_CharSet : public wxCharacterSet
{
public:
  iconv_t m2w, w2m;
  IC_CharSet(const wxChar*name) : wxCharacterSet(name), m2w((iconv_t)-1), w2m((iconv_t)-1) {}
  ~IC_CharSet() {
    if (m2w!=(iconv_t)-1) iconv_close(m2w);
    if (w2m!=(iconv_t)-1) iconv_close(w2m);
  }
  void LoadM2W() { if (m2w==(iconv_t)-1) m2w=iconv_open(WC_NAME,wxConvLibc.cWX2MB(cname)); }
  void LoadW2M() { if (w2m==(iconv_t)-1) w2m=iconv_open(wxConvLibc.cWX2MB(cname),WC_NAME); }
  size_t MB2WC(wchar_t*buf, const char*psz, size_t n) {
    LoadM2W();
    size_t inbuf = strlen(psz);
    size_t outbuf = n*SIZEOF_WCHAR_T;
    size_t res, cres;
    fprintf(stderr,"IC Convert to WC using %s\n",(const char*)wxConvLibc.cWX2MB(cname));
    if (buf) {
      // have destination buffer, convert there
      cres = iconv(m2w,&psz,&inbuf,(char**)&buf,&outbuf);
      res = n-(outbuf/SIZEOF_WCHAR_T);
      // convert to native endianness
      WC_BSWAP(buf, res)
    } else {
      // no destination buffer... convert using temp buffer
      // to calculate destination buffer requirement
      wchar_t tbuf[8];
      res = 0;
      do {
        buf = tbuf; outbuf = 8*SIZEOF_WCHAR_T;
        cres = iconv(m2w,&psz,&inbuf,(char**)&buf,&outbuf);
        res += 8-(outbuf/SIZEOF_WCHAR_T);
      } while ((cres==(size_t)-1) && (errno==E2BIG));
    }
    if (cres==(size_t)-1) return (size_t)-1;
    return res;
  }
  size_t WC2MB(char*buf, const wchar_t*psz, size_t n) {
    LoadW2M();
#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530)
    size_t inbuf = std::wcslen(psz);
#else
    size_t inbuf = ::wcslen(psz);
#endif
    size_t outbuf = n;
    size_t res, cres;
    fprintf(stderr,"IC Convert from WC using %s\n",(const char*)wxConvLibc.cWX2MB(cname));
#ifdef WC_NEED_BSWAP
    // 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)
    wchar_t*tmpbuf=(wchar_t*)malloc((inbuf+1)*SIZEOF_WCHAR_T);
    memcpy(tmpbuf,psz,(inbuf+1)*SIZEOF_WCHAR_T);
    WC_BSWAP(tmpbuf, inbuf)
    psz=tmpbuf;
#endif
    if (buf) {
      // have destination buffer, convert there
      cres = iconv(w2m,(const char**)&psz,&inbuf,&buf,&outbuf);
      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,(const char**)&psz,&inbuf,&buf,&outbuf);
        res += 16 - outbuf;
      } while ((cres==(size_t)-1) && (errno==E2BIG));
    }
#ifdef WC_NEED_BSWAP
    free(tmpbuf);
#endif
    if (cres==(size_t)-1) return (size_t)-1;
    return res;
  }
  bool usable() { return TRUE; }
};
#endif

#ifdef __WIN32__
class CP_CharSet : public wxCharacterSet
{
public:
  long CodePage;
  CP_CharSet(const wxChar*name) : wxCharacterSet(name), CodePage(CharsetToCodepage(name)) {}
  size_t MB2WC(wchar_t*buf, const char*psz, size_t n) {
    size_t len = MultiByteToWideChar(CodePage,0,psz,-1,buf,buf?n:0);
    return len?len:(size_t)-1;
  }
  size_t WC2MB(char*buf, const wchar_t*psz, size_t n) {
    size_t len = WideCharToMultiByte(CodePage,0,psz,-1,buf,buf?n:0,NULL,NULL);
    return len?len:(size_t)-1;
  }
  bool usable() { return CodePage!=-1; }
};
#endif

class EC_CharSet : public wxCharacterSet
{
public:
  // temporarily just use wxEncodingConverter stuff,
  // so that it works while a better implementation is built
  wxFontEncoding enc;
  wxEncodingConverter m2w, w2m;
  EC_CharSet(const wxChar*name) : wxCharacterSet(name), enc(wxFONTENCODING_SYSTEM)
  {
    if (name) enc = wxTheFontMapper->CharsetToEncoding(name, FALSE);
    m2w.Init(enc, wxFONTENCODING_UNICODE);
    w2m.Init(wxFONTENCODING_UNICODE, enc);
  }
  size_t MB2WC(wchar_t*buf, const char*psz, size_t n) {
    size_t inbuf = strlen(psz);
    fprintf(stderr,"EC Convert to WC using %d\n",enc);
    if (buf) m2w.Convert(psz,buf);
    return inbuf;
  }
  size_t WC2MB(char*buf, const wchar_t*psz, size_t n) {
#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530)
    size_t inbuf = std::wcslen(psz);
#else
    size_t inbuf = ::wcslen(psz);
#endif
    fprintf(stderr,"EC Convert from WC using %d\n",enc);
    if (buf) w2m.Convert(psz,buf);
    return inbuf;
  }
  bool usable() { return (enc!=wxFONTENCODING_SYSTEM) && (enc!=wxFONTENCODING_DEFAULT); }
};

static wxCharacterSet *wxGetCharacterSet(const wxChar *name)
{
  wxCharacterSet *cset = NULL;
  if (name) {
    if (!wxStricmp(name, wxT("UTF8")) || !wxStricmp(name, wxT("UTF-8"))) {
      cset = new ID_CharSet(name, &wxConvUTF8);
    } else {
#ifdef HAVE_ICONV_H
      cset = new IC_CharSet(name); // may not take NULL
#endif
    }
  }
  if (cset && cset->usable()) return cset;
  if (cset) delete cset;
#ifdef __WIN32__
  cset = new CP_CharSet(name); // may take NULL
  if (cset->usable()) return cset;
#endif
  if (cset) delete cset;
  cset = new EC_CharSet(name);
  if (cset->usable()) return cset;
  delete cset;
  return NULL;
}

wxCSConv::wxCSConv(const wxChar *charset)
{
  m_name = (wxChar *) NULL;
  m_cset = (wxCharacterSet *) NULL;
  m_deferred = TRUE;
  SetName(charset);
}

wxCSConv::~wxCSConv()
{
  if (m_name) free(m_name);
  if (m_cset) delete m_cset;
}

void wxCSConv::SetName(const wxChar *charset)
{
  if (charset) {
    m_name = wxStrdup(charset);
    m_deferred = TRUE;
  }
}

void wxCSConv::LoadNow()
{
//  wxPrintf(wxT("Conversion request\n"));
  if (m_deferred) {
    if (!m_name) {
#ifdef __UNIX__
#if defined(HAVE_LANGINFO_H) && defined(CODESET)
      // GNU libc provides current character set this way
      char*alang = nl_langinfo(CODESET);
      if (alang) SetName(wxConvLibc.cMB2WX(alang));
      else
#endif
      // if we can't get at the character set directly,
      // try to see if it's in the environment variables
      // (in most cases this won't work, but I was out of ideas)
      {
        wxChar *lang = wxGetenv(wxT("LC_ALL"));
        if (!lang) lang = wxGetenv(wxT("LC_CTYPE"));
        if (!lang) lang = wxGetenv(wxT("LANG"));
        wxChar *dot = lang ? wxStrchr(lang, wxT('.')) : (wxChar *)NULL;
        if (dot) SetName(dot+1);
      }
#endif
    }
    m_cset = wxGetCharacterSet(m_name);
    m_deferred = FALSE;
  }
}

size_t wxCSConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const
{
  ((wxCSConv *)this)->LoadNow(); // discard constness
  if (m_cset)
    return m_cset->MB2WC(buf, psz, n);

  // latin-1 (direct)
  size_t len=strlen(psz);
  if (buf) {
    for (size_t c=0; c<=len; c++)
      buf[c] = (unsigned char)(psz[c]);
  }
  return len;
}

size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const
{
  ((wxCSConv *)this)->LoadNow(); // discard constness
  if (m_cset)
    return m_cset->WC2MB(buf, psz, n);

  // latin-1 (direct)
#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530)
  size_t len=std::wcslen(psz);
#else
  size_t len=::wcslen(psz);
#endif
  if (buf) {
    for (size_t c=0; c<=len; c++)
      buf[c] = (psz[c]>0xff) ? '?' : psz[c];
  }
  return len;
}

#ifdef HAVE_ICONV_H
class IC_CharSetConverter
{
public:
  iconv_t cnv;
  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;
    size_t res = iconv(cnv,&psz,&inbuf,&buf,&outbuf);
    if (res==(size_t)-1) return (size_t)-1;
    return n-outbuf;
  }
};
#endif

class EC_CharSetConverter
{
public:
  wxEncodingConverter cnv;
  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;
  }
};

#else // !wxUSE_WCHAR_T

// ----------------------------------------------------------------------------
// stand-ins in absence of wchar_t
// ----------------------------------------------------------------------------

WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile;

#endif // wxUSE_WCHAR_T
Commit	Line	Data
6001e347 RR	1	/////////////////////////////////////////////////////////////////////////////
	2	// Name: strconv.cpp
	3	// Purpose: Unicode conversion classes
	4	// Author: Ove Kaaven, Robert Roebling, Vadim Zeitlin
	5	// Modified by:
	6	// Created: 29/01/98
	7	// RCS-ID: $Id$
	8	// Copyright: (c) 1999 Ove Kaaven, Robert Roebling, Vadim Zeitlin
	9	// Licence: wxWindows license
	10	/////////////////////////////////////////////////////////////////////////////
	11
f6bcfd97 BP	12	// ============================================================================
	13	// declarations
	14	// ============================================================================
	15
	16	// ----------------------------------------------------------------------------
	17	// headers
	18	// ----------------------------------------------------------------------------
	19
6001e347 RR	20	#ifdef __GNUG__
	21	#pragma implementation "strconv.h"
	22	#endif
	23
	24	// For compilers that support precompilation, includes "wx.h".
	25	#include "wx/wxprec.h"
	26
	27	#ifdef __BORLANDC__
	28	#pragma hdrstop
	29	#endif
	30
0a1c1e62 GRG	31	#ifdef __WXMSW__
	32	#include "wx/msw/private.h"
	33	#endif
	34
1cd52418	35	#include <errno.h>
6001e347 RR	36	#include <ctype.h>
	37	#include <string.h>
	38	#include <stdlib.h>
	39
	40	#ifdef __SALFORDC__
	41	#include <clib.h>
	42	#endif
	43
1cd52418 OK	44	#ifdef HAVE_ICONV_H
	45	#include <iconv.h>
	46	#endif
	47	#ifdef HAVE_LANGINFO_H
	48	#include <langinfo.h>
	49	#endif
	50
3e61dfb0 OK	51	#ifdef __WXMSW__
	52	#include <windows.h>
	53	#endif
	54
6001e347 RR	55	#include "wx/debug.h"
	56	#include "wx/strconv.h"
	57
d43088ee	58	#if defined(WORDS_BIGENDIAN) \|\| defined(__STDC_ISO_10646__)
1cd52418 OK	59	#define BSWAP_UCS4(str, len)
	60	#define BSWAP_UCS2(str, len)
	61	#else
	62	#define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT32_SWAP_ALWAYS(str[_c]); }
	63	#define BSWAP_UCS2(str, len) { unsigned _c; for (_c=0; _c<len; _c++) str[_c]=wxUINT16_SWAP_ALWAYS(str[_c]); }
	64	#define WC_NEED_BSWAP
	65	#endif
	66	#define BSWAP_UTF32(str, len) BSWAP_UCS4(str, len)
	67	#define BSWAP_UTF16(str, len) BSWAP_UCS2(str, len)
	68
	69	#if SIZEOF_WCHAR_T == 4
	70	#define WC_NAME "UCS4"
	71	#define WC_BSWAP BSWAP_UCS4
	72	#elif SIZEOF_WCHAR_T == 2
	73	#define WC_NAME "UTF16"
	74	#define WC_BSWAP BSWAP_UTF16
	75	#define WC_UTF16
	76	#endif
	77
f6bcfd97 BP	78	// ----------------------------------------------------------------------------
	79	// globals
	80	// ----------------------------------------------------------------------------
6001e347 RR	81
	82	WXDLLEXPORT_DATA(wxMBConv *) wxConvCurrent = &wxConvLibc;
	83
f6bcfd97 BP	84	// ============================================================================
	85	// implementation
	86	// ============================================================================
6001e347	87
f6bcfd97	88	#if wxUSE_WCHAR_T
6001e347	89
b0a6bb75 VZ	90	#ifdef WC_UTF16
b0a6bb75 VZ	91
1cd52418 OK	92	static size_t encode_utf16(wxUint32 input,wxUint16*output)
	93	{
	94	if (input<=0xffff) {
	95	if (output) *output++ = input;
	96	return 1;
	97	} else
	98	if (input>=0x110000) {
	99	return (size_t)-1;
	100	} else {
	101	if (output) {
	102	*output++ = (input >> 10)+0xd7c0;
	103	*output++ = (input&0x3ff)+0xdc00;
	104	}
	105	return 2;
	106	}
	107	}
	108
	109	static size_t decode_utf16(wxUint16*input,wxUint32&output)
	110	{
	111	if ((input<0xd800) \|\| (input>0xdfff)) {
	112	output = *input;
	113	return 1;
	114	} else
	115	if ((input[1]<0xdc00) \|\| (input[1]>=0xdfff)) {
	116	output = *input;
	117	return (size_t)-1;
	118	} else {
	119	output = ((input[0] - 0xd7c0) << 10) + (input[1] - 0xdc00);
	120	return 2;
	121	}
	122	}
	123
b0a6bb75 VZ	124	#endif // WC_UTF16
b0a6bb75 VZ	125
f6bcfd97	126	// ----------------------------------------------------------------------------
6001e347	127	// wxMBConv
f6bcfd97	128	// ----------------------------------------------------------------------------
6001e347 RR	129
	130	WXDLLEXPORT_DATA(wxMBConv) wxConvLibc;
	131
	132	size_t wxMBConv::MB2WC(wchar_t buf, const char psz, size_t n) const
	133	{
	134	return wxMB2WC(buf, psz, n);
	135	}
	136
	137	size_t wxMBConv::WC2MB(char buf, const wchar_t psz, size_t n) const
	138	{
	139	return wxWC2MB(buf, psz, n);
	140	}
	141
	142	const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const
	143	{
f6bcfd97	144	if (psz)
6001e347 RR	145	{
6001e347 RR	146	size_t nLen = MB2WC((wchar_t *) NULL, psz, 0);
f6bcfd97 BP	147	if (nLen == (size_t)-1)
f6bcfd97 BP	148	return wxWCharBuffer((wchar_t *) NULL);
6001e347 RR	149	wxWCharBuffer buf(nLen);
	150	MB2WC((wchar_t )(const wchar_t ) buf, psz, nLen);
	151	return buf;
f6bcfd97 BP	152	}
f6bcfd97 BP	153	else
6001e347 RR	154	return wxWCharBuffer((wchar_t *) NULL);
	155	}
	156
	157	const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *psz) const
	158	{
f6bcfd97	159	if (psz)
6001e347 RR	160	{
6001e347 RR	161	size_t nLen = WC2MB((char *) NULL, psz, 0);
f6bcfd97 BP	162	if (nLen == (size_t)-1)
f6bcfd97 BP	163	return wxCharBuffer((char *) NULL);
6001e347 RR	164	wxCharBuffer buf(nLen);
	165	WC2MB((char )(const char ) buf, psz, nLen);
	166	return buf;
f6bcfd97 BP	167	}
f6bcfd97 BP	168	else
6001e347 RR	169	return wxCharBuffer((char *) NULL);
	170	}
	171
f6bcfd97	172	// ----------------------------------------------------------------------------
6001e347	173	// standard file conversion
f6bcfd97	174	// ----------------------------------------------------------------------------
6001e347 RR	175
	176	WXDLLEXPORT_DATA(wxMBConvFile) wxConvFile;
	177
	178	// just use the libc conversion for now
	179	size_t wxMBConvFile::MB2WC(wchar_t buf, const char psz, size_t n) const
	180	{
	181	return wxMB2WC(buf, psz, n);
	182	}
	183
	184	size_t wxMBConvFile::WC2MB(char buf, const wchar_t psz, size_t n) const
	185	{
	186	return wxWC2MB(buf, psz, n);
	187	}
	188
f6bcfd97	189	// ----------------------------------------------------------------------------
6001e347	190	// standard gdk conversion
f6bcfd97 BP	191	// ----------------------------------------------------------------------------
	192
	193	#ifdef __WXGTK12__
6001e347 RR	194
	195	WXDLLEXPORT_DATA(wxMBConvGdk) wxConvGdk;
	196
	197	#include <gdk/gdk.h>
	198
	199	size_t wxMBConvGdk::MB2WC(wchar_t buf, const char psz, size_t n) const
	200	{
	201	if (buf) {
	202	return gdk_mbstowcs((GdkWChar *)buf, psz, n);
	203	} else {
	204	GdkWChar *nbuf = new GdkWChar[n=strlen(psz)];
	205	size_t len = gdk_mbstowcs(nbuf, psz, n);
	206	delete [] nbuf;
	207	return len;
	208	}
	209	}
	210
	211	size_t wxMBConvGdk::WC2MB(char buf, const wchar_t psz, size_t n) const
	212	{
	213	char mbstr = gdk_wcstombs((GdkWChar )psz);
	214	size_t len = mbstr ? strlen(mbstr) : 0;
	215	if (buf) {
	216	if (len > n) len = n;
	217	memcpy(buf, psz, len);
	218	if (len < n) buf[len] = 0;
	219	}
	220	return len;
	221	}
f6bcfd97	222
6001e347 RR	223	#endif // GTK > 1.0
	224
	225	// ----------------------------------------------------------------------------
	226	// UTF-7
	227	// ----------------------------------------------------------------------------
	228
	229	WXDLLEXPORT_DATA(wxMBConvUTF7) wxConvUTF7;
	230
	231	#if 0
	232	static char utf7_setD[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ"
	233	"abcdefghijklmnopqrstuvwxyz"
	234	"0123456789'(),-./:?";
	235	static char utf7_setO[]="!\"#$%&*;<=>@[]^_`{\|}";
	236	static char utf7_setB[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ"
	237	"abcdefghijklmnopqrstuvwxyz"
	238	"0123456789+/";
	239	#endif
	240
	241	// TODO: write actual implementations of UTF-7 here
	242	size_t wxMBConvUTF7::MB2WC(wchar_t * WXUNUSED(buf),
	243	const char * WXUNUSED(psz),
	244	size_t WXUNUSED(n)) const
	245	{
	246	return 0;
	247	}
	248
	249	size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf),
	250	const wchar_t * WXUNUSED(psz),
	251	size_t WXUNUSED(n)) const
	252	{
	253	return 0;
	254	}
	255
f6bcfd97	256	// ----------------------------------------------------------------------------
6001e347	257	// UTF-8
f6bcfd97	258	// ----------------------------------------------------------------------------
6001e347 RR	259
	260	WXDLLEXPORT_DATA(wxMBConvUTF8) wxConvUTF8;
	261
1cd52418	262	static wxUint32 utf8_max[]={0x7f,0x7ff,0xffff,0x1fffff,0x3ffffff,0x7fffffff,0xffffffff};
6001e347 RR	263
	264	size_t wxMBConvUTF8::MB2WC(wchar_t buf, const char psz, size_t n) const
	265	{
	266	size_t len = 0;
	267
	268	while (*psz && ((!buf) \|\| (len<n))) {
	269	unsigned char cc=*psz++, fc=cc;
	270	unsigned cnt;
	271	for (cnt=0; fc&0x80; cnt++) fc<<=1;
	272	if (!cnt) {
	273	// plain ASCII char
	274	if (buf) *buf++=cc;
	275	len++;
	276	} else {
	277	cnt--;
	278	if (!cnt) {
	279	// invalid UTF-8 sequence
	280	return (size_t)-1;
	281	} else {
	282	unsigned ocnt=cnt-1;
1cd52418	283	wxUint32 res=cc&(0x3f>>cnt);
6001e347 RR	284	while (cnt--) {
	285	cc = *psz++;
	286	if ((cc&0xC0)!=0x80) {
	287	// invalid UTF-8 sequence
	288	return (size_t)-1;
	289	}
	290	res=(res<<6)\|(cc&0x3f);
	291	}
	292	if (res<=utf8_max[ocnt]) {
	293	// illegal UTF-8 encoding
	294	return (size_t)-1;
	295	}
1cd52418 OK	296	#ifdef WC_UTF16
	297	size_t pa = encode_utf16(res, buf);
	298	if (pa == (size_t)-1)
	299	return (size_t)-1;
	300	if (buf) buf+=pa;
	301	len+=pa;
	302	#else
6001e347 RR	303	if (buf) *buf++=res;
6001e347 RR	304	len++;
1cd52418	305	#endif
6001e347 RR	306	}
	307	}
	308	}
	309	if (buf && (len<n)) *buf = 0;
	310	return len;
	311	}
	312
	313	size_t wxMBConvUTF8::WC2MB(char buf, const wchar_t psz, size_t n) const
	314	{
	315	size_t len = 0;
	316
	317	while (*psz && ((!buf) \|\| (len<n))) {
1cd52418 OK	318	wxUint32 cc;
	319	#ifdef WC_UTF16
	320	size_t pa = decode_utf16(psz,cc);
	321	psz += (pa == (size_t)-1) ? 1 : pa;
	322	#else
	323	cc=(*psz++)&0x7fffffff;
	324	#endif
6001e347 RR	325	unsigned cnt;
	326	for (cnt=0; cc>utf8_max[cnt]; cnt++);
	327	if (!cnt) {
	328	// plain ASCII char
	329	if (buf) *buf++=cc;
	330	len++;
	331	} else {
	332	len+=cnt+1;
	333	if (buf) {
	334	buf++=(-128>>cnt)\|((cc>>(cnt6))&(0x3f>>cnt));
	335	while (cnt--)
	336	buf++=0x80\|((cc>>(cnt6))&0x3f);
	337	}
	338	}
	339	}
	340	if (buf && (len<n)) *buf = 0;
	341	return len;
	342	}
	343
	344	// ----------------------------------------------------------------------------
	345	// specified character set
	346	// ----------------------------------------------------------------------------
	347
f6bcfd97 BP	348	WXDLLEXPORT_DATA(wxCSConv) wxConvLocal((const wxChar *)NULL);
	349
	350	#include "wx/encconv.h"
	351	#include "wx/fontmap.h"
6001e347	352
1cd52418 OK	353	// TODO: add some tables here
	354	// - perhaps common encodings to common codepages (for Win32)
	355	// - perhaps common encodings to objects ("UTF8" -> wxConvUTF8)
	356	// - move wxEncodingConverter meat in here
	357
	358	#ifdef __WIN32__
	359	#include "wx/msw/registry.h"
	360	// this should work if M$ Internet Exploiter is installed
	361	static long CharsetToCodepage(const wxChar *name)
	362	{
	363	if (!name) return GetACP();
	364	long CP=-1;
	365	wxString cn(name);
	366	do {
	367	wxString path(wxT("MIME\\Database\\Charset\\"));
	368	path += cn;
	369	wxRegKey key(wxRegKey::HKCR,path);
	370
	371	/* two cases: either there's an AliasForCharset string,
	372	* or there are Codepage and InternetEncoding dwords.
	373	* The InternetEncoding gives us the actual encoding,
	374	* the Codepage just says which Windows character set to
	375	* use when displaying the data.
	376	*/
	377	if (key.QueryValue(wxT("InternetEncoding"),&CP)) break;
	378	// no encoding, see if it's an alias
	379	if (!key.QueryValue(wxT("AliasForCharset"),cn)) break;
	380	} while (1);
	381	return CP;
	382	}
	383	#endif
	384
6001e347 RR	385	class wxCharacterSet
6001e347 RR	386	{
1cd52418 OK	387	public:
	388	const wxChar*cname;
	389	wxCharacterSet(const wxChar*name) : cname(name) {}
	390	virtual ~wxCharacterSet() {}
	391	virtual size_t MB2WC(wchar_tbuf, const charpsz, size_t n) { return (size_t)-1; }
	392	virtual size_t WC2MB(charbuf, const wchar_tpsz, size_t n) { return (size_t)-1; }
	393	virtual bool usable() { return FALSE; }
	394	};
	395
	396	class ID_CharSet : public wxCharacterSet
	397	{
	398	public:
	399	wxMBConv*work;
	400	ID_CharSet(const wxCharname,wxMBConvcnv) : wxCharacterSet(name), work(cnv) {}
	401	size_t MB2WC(wchar_tbuf, const charpsz, size_t n)
	402	{ return work ? work->MB2WC(buf,psz,n) : (size_t)-1; }
	403	size_t WC2MB(charbuf, const wchar_tpsz, size_t n)
	404	{ return work ? work->WC2MB(buf,psz,n) : (size_t)-1; }
	405	bool usable() { return work!=NULL; }
	406	};
	407
	408	#ifdef HAVE_ICONV_H
	409	class IC_CharSet : public wxCharacterSet
	410	{
	411	public:
	412	iconv_t m2w, w2m;
	413	IC_CharSet(const wxChar*name) : wxCharacterSet(name), m2w((iconv_t)-1), w2m((iconv_t)-1) {}
	414	~IC_CharSet() {
	415	if (m2w!=(iconv_t)-1) iconv_close(m2w);
	416	if (w2m!=(iconv_t)-1) iconv_close(w2m);
	417	}
	418	void LoadM2W() { if (m2w==(iconv_t)-1) m2w=iconv_open(WC_NAME,wxConvLibc.cWX2MB(cname)); }
	419	void LoadW2M() { if (w2m==(iconv_t)-1) w2m=iconv_open(wxConvLibc.cWX2MB(cname),WC_NAME); }
	420	size_t MB2WC(wchar_tbuf, const charpsz, size_t n) {
	421	LoadM2W();
	422	size_t inbuf = strlen(psz);
	423	size_t outbuf = n*SIZEOF_WCHAR_T;
	424	size_t res, cres;
	425	fprintf(stderr,"IC Convert to WC using %s\n",(const char*)wxConvLibc.cWX2MB(cname));
	426	if (buf) {
	427	// have destination buffer, convert there
	428	cres = iconv(m2w,&psz,&inbuf,(char**)&buf,&outbuf);
	429	res = n-(outbuf/SIZEOF_WCHAR_T);
	430	// convert to native endianness
	431	WC_BSWAP(buf, res)
	432	} else {
	433	// no destination buffer... convert using temp buffer
	434	// to calculate destination buffer requirement
	435	wchar_t tbuf[8];
	436	res = 0;
	437	do {
	438	buf = tbuf; outbuf = 8*SIZEOF_WCHAR_T;
	439	cres = iconv(m2w,&psz,&inbuf,(char**)&buf,&outbuf);
	440	res += 8-(outbuf/SIZEOF_WCHAR_T);
	441	} while ((cres==(size_t)-1) && (errno==E2BIG));
	442	}
	443	if (cres==(size_t)-1) return (size_t)-1;
	444	return res;
	445	}
	446	size_t WC2MB(charbuf, const wchar_tpsz, size_t n) {
	447	LoadW2M();
	448	#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530)
	449	size_t inbuf = std::wcslen(psz);
	450	#else
451	size_t inbuf = ::wcslen(psz);
452	#endif
453	size_t outbuf = n;
454	size_t res, cres;
455	fprintf(stderr,"IC Convert from WC using %s\n",(const char*)wxConvLibc.cWX2MB(cname));
456	#ifdef WC_NEED_BSWAP
457	// need to copy to temp buffer to switch endianness
458	// this absolutely doesn't rock!
459	// (no, doing WC_BSWAP twice on the original buffer won't help, as it
460	// could be in read-only memory, or be accessed in some other thread)
461	wchar_ttmpbuf=(wchar_t)malloc((inbuf+1)*SIZEOF_WCHAR_T);
462	memcpy(tmpbuf,psz,(inbuf+1)*SIZEOF_WCHAR_T);
463	WC_BSWAP(tmpbuf, inbuf)
464	psz=tmpbuf;
465	#endif
466	if (buf) {
467	// have destination buffer, convert there
468	cres = iconv(w2m,(const char**)&psz,&inbuf,&buf,&outbuf);
469	res = n-outbuf;
470	} else {
471	// no destination buffer... convert using temp buffer
472	// to calculate destination buffer requirement
473	char tbuf[16];
474	res = 0;
475	do {
476	buf = tbuf; outbuf = 16;
477	cres = iconv(w2m,(const char**)&psz,&inbuf,&buf,&outbuf);
478	res += 16 - outbuf;
479	} while ((cres==(size_t)-1) && (errno==E2BIG));
480	}
481	#ifdef WC_NEED_BSWAP
482	free(tmpbuf);
483	#endif
484	if (cres==(size_t)-1) return (size_t)-1;
485	return res;
486	}
487	bool usable() { return TRUE; }
488	};
489	#endif
490
491	#ifdef __WIN32__
492	class CP_CharSet : public wxCharacterSet
493	{
494	public:
495	long CodePage;
496	CP_CharSet(const wxChar*name) : wxCharacterSet(name), CodePage(CharsetToCodepage(name)) {}
497	size_t MB2WC(wchar_tbuf, const charpsz, size_t n) {
498	size_t len = MultiByteToWideChar(CodePage,0,psz,-1,buf,buf?n:0);
499	return len?len:(size_t)-1;
500	}
501	size_t WC2MB(charbuf, const wchar_tpsz, size_t n) {
502	size_t len = WideCharToMultiByte(CodePage,0,psz,-1,buf,buf?n:0,NULL,NULL);
503	return len?len:(size_t)-1;
504	}
505	bool usable() { return CodePage!=-1; }
506	};
507	#endif
508
509	class EC_CharSet : public wxCharacterSet
510	{
6001e347	511	public:
f6bcfd97 BP	512	// temporarily just use wxEncodingConverter stuff,
	513	// so that it works while a better implementation is built
	514	wxFontEncoding enc;
	515	wxEncodingConverter m2w, w2m;
1cd52418	516	EC_CharSet(const wxChar*name) : wxCharacterSet(name), enc(wxFONTENCODING_SYSTEM)
f6bcfd97	517	{
1cd52418	518	if (name) enc = wxTheFontMapper->CharsetToEncoding(name, FALSE);
f6bcfd97 BP	519	m2w.Init(enc, wxFONTENCODING_UNICODE);
f6bcfd97 BP	520	w2m.Init(wxFONTENCODING_UNICODE, enc);
6001e347	521	}
1cd52418 OK	522	size_t MB2WC(wchar_tbuf, const charpsz, size_t n) {
	523	size_t inbuf = strlen(psz);
	524	fprintf(stderr,"EC Convert to WC using %d\n",enc);
	525	if (buf) m2w.Convert(psz,buf);
	526	return inbuf;
	527	}
	528	size_t WC2MB(charbuf, const wchar_tpsz, size_t n) {
	529	#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530)
	530	size_t inbuf = std::wcslen(psz);
	531	#else
	532	size_t inbuf = ::wcslen(psz);
	533	#endif
	534	fprintf(stderr,"EC Convert from WC using %d\n",enc);
	535	if (buf) w2m.Convert(psz,buf);
	536	return inbuf;
	537	}
	538	bool usable() { return (enc!=wxFONTENCODING_SYSTEM) && (enc!=wxFONTENCODING_DEFAULT); }
f6bcfd97	539	};
6001e347	540
f6bcfd97	541	static wxCharacterSet wxGetCharacterSet(const wxChar name)
6001e347	542	{
1cd52418 OK	543	wxCharacterSet *cset = NULL;
	544	if (name) {
	545	if (!wxStricmp(name, wxT("UTF8")) \|\| !wxStricmp(name, wxT("UTF-8"))) {
	546	cset = new ID_CharSet(name, &wxConvUTF8);
	547	} else {
	548	#ifdef HAVE_ICONV_H
	549	cset = new IC_CharSet(name); // may not take NULL
	550	#endif
	551	}
	552	}
	553	if (cset && cset->usable()) return cset;
	554	if (cset) delete cset;
	555	#ifdef __WIN32__
	556	cset = new CP_CharSet(name); // may take NULL
	557	if (cset->usable()) return cset;
	558	#endif
	559	if (cset) delete cset;
	560	cset = new EC_CharSet(name);
	561	if (cset->usable()) return cset;
	562	delete cset;
	563	return NULL;
6001e347 RR	564	}
6001e347 RR	565
6001e347 RR	566	wxCSConv::wxCSConv(const wxChar *charset)
	567	{
	568	m_name = (wxChar *) NULL;
	569	m_cset = (wxCharacterSet *) NULL;
	570	m_deferred = TRUE;
	571	SetName(charset);
	572	}
	573
	574	wxCSConv::~wxCSConv()
	575	{
	576	if (m_name) free(m_name);
f6bcfd97	577	if (m_cset) delete m_cset;
6001e347 RR	578	}
	579
	580	void wxCSConv::SetName(const wxChar *charset)
	581	{
	582	if (charset) {
f6bcfd97	583	m_name = wxStrdup(charset);
6001e347	584	m_deferred = TRUE;
6001e347 RR	585	}
	586	}
	587
	588	void wxCSConv::LoadNow()
	589	{
223d09f6	590	// wxPrintf(wxT("Conversion request\n"));
6001e347 RR	591	if (m_deferred) {
	592	if (!m_name) {
	593	#ifdef __UNIX__
1cd52418 OK	594	#if defined(HAVE_LANGINFO_H) && defined(CODESET)
	595	// GNU libc provides current character set this way
	596	char*alang = nl_langinfo(CODESET);
	597	if (alang) SetName(wxConvLibc.cMB2WX(alang));
	598	else
	599	#endif
	600	// if we can't get at the character set directly,
	601	// try to see if it's in the environment variables
	602	// (in most cases this won't work, but I was out of ideas)
	603	{
	604	wxChar *lang = wxGetenv(wxT("LC_ALL"));
	605	if (!lang) lang = wxGetenv(wxT("LC_CTYPE"));
	606	if (!lang) lang = wxGetenv(wxT("LANG"));
	607	wxChar dot = lang ? wxStrchr(lang, wxT('.')) : (wxChar )NULL;
	608	if (dot) SetName(dot+1);
	609	}
6001e347 RR	610	#endif
6001e347 RR	611	}
f6bcfd97	612	m_cset = wxGetCharacterSet(m_name);
6001e347 RR	613	m_deferred = FALSE;
	614	}
	615	}
	616
	617	size_t wxCSConv::MB2WC(wchar_t buf, const char psz, size_t n) const
	618	{
	619	((wxCSConv *)this)->LoadNow(); // discard constness
1cd52418 OK	620	if (m_cset)
	621	return m_cset->MB2WC(buf, psz, n);
	622
	623	// latin-1 (direct)
	624	size_t len=strlen(psz);
6001e347	625	if (buf) {
1cd52418 OK	626	for (size_t c=0; c<=len; c++)
1cd52418 OK	627	buf[c] = (unsigned char)(psz[c]);
6001e347	628	}
1cd52418	629	return len;
6001e347 RR	630	}
	631
	632	size_t wxCSConv::WC2MB(char buf, const wchar_t psz, size_t n) const
	633	{
	634	((wxCSConv *)this)->LoadNow(); // discard constness
1cd52418 OK	635	if (m_cset)
	636	return m_cset->WC2MB(buf, psz, n);
	637
	638	// latin-1 (direct)
d834f22c	639	#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530)
1cd52418	640	size_t len=std::wcslen(psz);
d834f22c	641	#else
1cd52418	642	size_t len=::wcslen(psz);
d834f22c	643	#endif
1cd52418 OK	644	if (buf) {
	645	for (size_t c=0; c<=len; c++)
	646	buf[c] = (psz[c]>0xff) ? '?' : psz[c];
	647	}
	648	return len;
6001e347 RR	649	}
6001e347 RR	650
1cd52418 OK	651	#ifdef HAVE_ICONV_H
	652	class IC_CharSetConverter
	653	{
	654	public:
	655	iconv_t cnv;
	656	IC_CharSetConverter(IC_CharSetfrom,IC_CharSetto) {
	657	cnv=iconv_open(wxConvLibc.cWX2MB(to->cname),wxConvLibc.cWX2MB(from->cname));
	658	}
	659	~IC_CharSetConverter() {
	660	if (cnv!=(iconv_t)-1) iconv_close(cnv);
	661	}
	662	size_t Convert(charbuf, const charpsz, size_t n) {
	663	size_t inbuf = strlen(psz);
	664	size_t outbuf = n;
	665	size_t res = iconv(cnv,&psz,&inbuf,&buf,&outbuf);
	666	if (res==(size_t)-1) return (size_t)-1;
	667	return n-outbuf;
	668	}
	669	};
	670	#endif
	671
	672	class EC_CharSetConverter
	673	{
	674	public:
	675	wxEncodingConverter cnv;
	676	EC_CharSetConverter(EC_CharSetfrom,EC_CharSetto) {
	677	cnv.Init(from->enc,to->enc);
	678	}
	679	size_t Convert(charbuf, const charpsz, size_t n) {
	680	size_t inbuf = strlen(psz);
	681	if (buf) cnv.Convert(psz,buf);
	682	return inbuf;
	683	}
	684	};
	685
f6bcfd97 BP	686	#else // !wxUSE_WCHAR_T
	687
	688	// ----------------------------------------------------------------------------
	689	// stand-ins in absence of wchar_t
	690	// ----------------------------------------------------------------------------
	691
	692	WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile;
	693
	694	#endif // wxUSE_WCHAR_T
6001e347 RR	695
6001e347 RR	696