X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/fc7a2a602b4bd9308be21bc07e40a680f483438e..fef5c55604e20e2a90843a2278b1f5c8c817dcee:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 2279100d93..ea5e74110f 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -5,7 +5,8 @@ // 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 ///////////////////////////////////////////////////////////////////////////// @@ -17,7 +18,7 @@ // headers // ---------------------------------------------------------------------------- -#ifdef __GNUG__ +#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA) #pragma implementation "strconv.h" #endif @@ -33,6 +34,10 @@ #include "wx/log.h" #endif // WX_PRECOMP +#include "wx/strconv.h" + +#if wxUSE_WCHAR_T + #ifdef __WXMSW__ #include "wx/msw/private.h" #endif @@ -45,53 +50,14 @@ #include #include -#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 #endif @@ -103,6 +69,13 @@ IMPLEMENT_DYNAMIC_CLASS(wxStrConvModule, wxModule) #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 // ---------------------------------------------------------------------------- @@ -110,12 +83,6 @@ IMPLEMENT_DYNAMIC_CLASS(wxStrConvModule, wxModule) #define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c=0x110000) @@ -163,14 +130,14 @@ static size_t encode_utf16(wxUint32 input, wchar_t *output) { 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)) { @@ -189,55 +156,16 @@ static size_t decode_utf16(const wchar_t* input, wxUint32& output) } } -#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 ) @@ -266,8 +194,8 @@ const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *pwz) const 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; } @@ -279,10 +207,22 @@ const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *pwz) const } // ---------------------------------------------------------------------------- -// 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" @@ -313,8 +253,6 @@ size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf), // UTF-8 // ---------------------------------------------------------------------------- -WXDLLIMPEXP_DATA_BASE(wxMBConvUTF8) wxConvUTF8; - static wxUint32 utf8_max[]= { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff, 0xffffffff }; @@ -363,7 +301,8 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const 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) @@ -390,7 +329,8 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const { 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; @@ -422,47 +362,472 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const 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 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 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 1) + *buf++ = cc[1]; + } + len += pa; + psz += sizeof(wxUint32); + } + + if (buf && lenCharsetToEncoding(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 @@ -806,22 +1301,36 @@ public: #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) @@ -829,7 +1338,7 @@ public: 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) @@ -838,102 +1347,55 @@ public: 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(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 -// ---------------------------------------------------------------------------- +// ============================================================================ +// 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__) && !defined(__WXUNIVERSAL__) - cset = new CP_CharSet(name); - if ( cset->usable() ) - return cset; + Init(); - delete cset; - cset = NULL; -#endif // defined(__WIN32__) && !defined(__WXMICROWIN__) && !defined(__WXUNIVERSAL__) - -#if wxUSE_FONTMAP - cset = new EC_CharSet(name); - if ( cset->usable() ) - return cset; - - 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() @@ -942,27 +1404,31 @@ 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) @@ -970,39 +1436,175 @@ 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 ) + // 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); @@ -1018,22 +1620,65 @@ size_t wxCSConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const 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