X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/fd5e476dbb9245a9ec755eeaa5b0e4b4556d2f8c..b9efe021b554fa3967d1442cf758435c5cd5ae8f:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index dfd5295d29..aea1e1a18a 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -70,6 +70,7 @@ #ifdef HAVE_ICONV #include + #include "wx/thread.h" #endif #include "wx/encconv.h" @@ -151,7 +152,7 @@ static size_t decode_utf16(const wxUint16* input, wxUint32& output) output = *input; return 1; } - else if ((input[1]<0xdc00) || (input[1]>=0xdfff)) + else if ((input[1]<0xdc00) || (input[1]>0xdfff)) { output = *input; return (size_t)-1; @@ -217,14 +218,18 @@ const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *pwz) const return buf; } -size_t wxMBConv::MB2WC(wchar_t* szBuffer, const char* szString, - size_t outsize, size_t nStringLen) const +const wxWCharBuffer wxMBConv::cMB2WC(const char *szString, size_t nStringLen, size_t* pOutSize) const { + wxASSERT(pOutSize != NULL); + const char* szEnd = szString + nStringLen + 1; const char* szPos = szString; const char* szStart = szPos; size_t nActualLength = 0; + size_t nCurrentSize = nStringLen; //try normal size first (should never resize?) + + wxWCharBuffer theBuffer(nCurrentSize); //Convert the string until the length() is reached, continuing the //loop every time a null character is reached @@ -237,38 +242,57 @@ size_t wxMBConv::MB2WC(wchar_t* szBuffer, const char* szString, //Invalid conversion? if( nLen == (size_t)-1 ) - return nLen; + { + *pOutSize = 0; + theBuffer.data()[0u] = wxT('\0'); + return theBuffer; + } + //Increase the actual length (+1 for current null character) nActualLength += nLen + 1; - //Only copy data in if buffer size is big enough - if (szBuffer != NULL && - nActualLength <= outsize) + //if buffer too big, realloc the buffer + if (nActualLength > (nCurrentSize+1)) { - //Convert the current (sub)string - if ( MB2WC(&szBuffer[szPos - szStart], szPos, nLen + 1) == (size_t)-1 ) - return (size_t)-1; + wxWCharBuffer theNewBuffer(nCurrentSize << 1); + memcpy(theNewBuffer.data(), theBuffer.data(), nCurrentSize * sizeof(wchar_t)); + theBuffer = theNewBuffer; + nCurrentSize <<= 1; + } + + //Convert the current (sub)string + if ( MB2WC(&theBuffer.data()[szPos - szStart], szPos, nLen + 1) == (size_t)-1 ) + { + *pOutSize = 0; + theBuffer.data()[0u] = wxT('\0'); + return theBuffer; } //Increment to next (sub)string - //Note that we have to use strlen here instead of nLen - //here because XX2XX gives us the size of the output buffer, - //not neccessarly the length of the string + //Note that we have to use strlen instead of nLen here + //because XX2XX gives us the size of the output buffer, + //which is not necessarily the length of the string szPos += strlen(szPos) + 1; } - return nActualLength - 1; //success - return actual length + //success - return actual length and the buffer + *pOutSize = nActualLength; + return theBuffer; } -size_t wxMBConv::WC2MB(char* szBuffer, const wchar_t* szString, - size_t outsize, size_t nStringLen) const +const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *szString, size_t nStringLen, size_t* pOutSize) const { + wxASSERT(pOutSize != NULL); + const wchar_t* szEnd = szString + nStringLen + 1; const wchar_t* szPos = szString; const wchar_t* szStart = szPos; size_t nActualLength = 0; + size_t nCurrentSize = nStringLen << 2; //try * 4 first + + wxCharBuffer theBuffer(nCurrentSize); //Convert the string until the length() is reached, continuing the //loop every time a null character is reached @@ -281,28 +305,42 @@ size_t wxMBConv::WC2MB(char* szBuffer, const wchar_t* szString, //Invalid conversion? if( nLen == (size_t)-1 ) - return nLen; + { + *pOutSize = 0; + theBuffer.data()[0u] = wxT('\0'); + return theBuffer; + } //Increase the actual length (+1 for current null character) nActualLength += nLen + 1; - - //Only copy data in if buffer size is big enough - if (szBuffer != NULL && - nActualLength <= outsize) + + //if buffer too big, realloc the buffer + if (nActualLength > (nCurrentSize+1)) { - //Convert the current (sub)string - if(WC2MB(&szBuffer[szPos - szStart], szPos, nLen + 1) == (size_t)-1 ) - return (size_t)-1; + wxCharBuffer theNewBuffer(nCurrentSize << 1); + memcpy(theNewBuffer.data(), theBuffer.data(), nCurrentSize); + theBuffer = theNewBuffer; + nCurrentSize <<= 1; + } + + //Convert the current (sub)string + if(WC2MB(&theBuffer.data()[szPos - szStart], szPos, nLen + 1) == (size_t)-1 ) + { + *pOutSize = 0; + theBuffer.data()[0u] = wxT('\0'); + return theBuffer; } //Increment to next (sub)string - //Note that we have to use wxWcslen here instead of nLen - //here because XX2XX gives us the size of the output buffer, - //not neccessarly the length of the string + //Note that we have to use wxWcslen instead of nLen here + //because XX2XX gives us the size of the output buffer, + //which is not necessarily the length of the string szPos += wxWcslen(szPos) + 1; } - return nActualLength - 1; //success - return actual length + //success - return actual length and the buffer + *pOutSize = nActualLength; + return theBuffer; } // ---------------------------------------------------------------------------- @@ -318,8 +356,42 @@ size_t wxMBConvLibc::WC2MB(char *buf, const wchar_t *psz, size_t n) const { return wxWC2MB(buf, psz, n); } + +#ifdef __UNIX__ + +// ---------------------------------------------------------------------------- +// wxConvBrokenFileNames +// ---------------------------------------------------------------------------- + +wxConvBrokenFileNames::wxConvBrokenFileNames(const wxChar *charset) +{ + if ( !charset || wxStricmp(charset, _T("UTF-8")) == 0 + || wxStricmp(charset, _T("UTF8")) == 0 ) + m_conv = new wxMBConvUTF8(wxMBConvUTF8::MAP_INVALID_UTF8_TO_OCTAL); + else + m_conv = new wxCSConv(charset); +} + +size_t +wxConvBrokenFileNames::MB2WC(wchar_t *outputBuf, + const char *psz, + size_t outputSize) const +{ + return m_conv->MB2WC( outputBuf, psz, outputSize ); +} + +size_t +wxConvBrokenFileNames::WC2MB(char *outputBuf, + const wchar_t *psz, + size_t outputSize) const +{ + return m_conv->WC2MB( outputBuf, psz, outputSize ); +} + +#endif + // ---------------------------------------------------------------------------- -// UTF-7 +// UTF-7 // ---------------------------------------------------------------------------- // Implementation (C) 2004 Fredrik Roubert @@ -365,7 +437,6 @@ static const unsigned char utf7unb64[] = size_t wxMBConvUTF7::MB2WC(wchar_t *buf, const char *psz, size_t n) const { - size_t len = 0; while (*psz && ((!buf) || (len < n))) @@ -399,7 +470,7 @@ size_t wxMBConvUTF7::MB2WC(wchar_t *buf, const char *psz, size_t n) const d += cc; for (l += 6; l >= 8; lsb = !lsb) { - c = (d >> (l -= 8)) % 256; + c = (unsigned char)((d >> (l -= 8)) % 256); if (lsb) { if (buf) @@ -408,7 +479,7 @@ size_t wxMBConvUTF7::MB2WC(wchar_t *buf, const char *psz, size_t n) const } else if (buf) - *buf = c << 8; + *buf = (wchar_t)(c << 8); } } if (*psz == '-') @@ -455,8 +526,7 @@ static const unsigned char utf7encode[128] = 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 3, 3 }; -size_t wxMBConvUTF7::WC2MB(char *buf, const wchar_t -*psz, size_t n) const +size_t wxMBConvUTF7::WC2MB(char *buf, const wchar_t *psz, size_t n) const { @@ -473,7 +543,7 @@ size_t wxMBConvUTF7::WC2MB(char *buf, const wchar_t len++; } #ifndef WC_UTF16 - else if (cc > ((const wchar_t)0xffff)) + else if (((wxUint32)cc) > 0xffff) { // no surrogate pair generation (yet?) return (size_t)-1; @@ -531,12 +601,19 @@ size_t wxMBConvUTF7::WC2MB(char *buf, const wchar_t static wxUint32 utf8_max[]= { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff, 0xffffffff }; +// boundaries of the private use area we use to (temporarily) remap invalid +// characters invalid in a UTF-8 encoded string +const wxUint32 wxUnicodePUA = 0x100000; +const wxUint32 wxUnicodePUAEnd = wxUnicodePUA + 256; + size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const { size_t len = 0; while (*psz && ((!buf) || (len < n))) { + const char *opsz = psz; + bool invalid = false; unsigned char cc = *psz++, fc = cc; unsigned cnt; for (cnt = 0; fc & 0x80; cnt++) @@ -547,6 +624,15 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const if (buf) *buf++ = cc; len++; + + // escape the escape character for octal escapes + if ((m_options & MAP_INVALID_UTF8_TO_OCTAL) + && cc == '\\' && (!buf || len < n)) + { + if (buf) + *buf++ = cc; + len++; + } } else { @@ -554,7 +640,7 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const if (!cnt) { // invalid UTF-8 sequence - return (size_t)-1; + invalid = true; } else { @@ -562,32 +648,93 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const wxUint32 res = cc & (0x3f >> cnt); while (cnt--) { - cc = *psz++; + cc = *psz; if ((cc & 0xC0) != 0x80) { // invalid UTF-8 sequence - return (size_t)-1; + invalid = true; + break; } + psz++; res = (res << 6) | (cc & 0x3f); } - if (res <= utf8_max[ocnt]) + if (invalid || res <= utf8_max[ocnt]) { // illegal UTF-8 encoding - return (size_t)-1; + invalid = true; + } + else if ((m_options & MAP_INVALID_UTF8_TO_PUA) && + res >= wxUnicodePUA && res < wxUnicodePUAEnd) + { + // if one of our PUA characters turns up externally + // it must also be treated as an illegal sequence + // (a bit like you have to escape an escape character) + invalid = true; } + else + { #ifdef WC_UTF16 - // 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) - buf += pa; - len += pa; + // cast is ok because wchar_t == wxUuint16 if WC_UTF16 + size_t pa = encode_utf16(res, (wxUint16 *)buf); + if (pa == (size_t)-1) + { + invalid = true; + } + else + { + if (buf) + buf += pa; + len += pa; + } #else // !WC_UTF16 - if (buf) - *buf++ = res; - len++; + if (buf) + *buf++ = res; + len++; #endif // WC_UTF16/!WC_UTF16 + } + } + if (invalid) + { + if (m_options & MAP_INVALID_UTF8_TO_PUA) + { + while (opsz < psz && (!buf || len < n)) + { +#ifdef WC_UTF16 + // cast is ok because wchar_t == wxUuint16 if WC_UTF16 + size_t pa = encode_utf16((unsigned char)*opsz + wxUnicodePUA, (wxUint16 *)buf); + wxASSERT(pa != (size_t)-1); + if (buf) + buf += pa; + opsz++; + len += pa; +#else + if (buf) + *buf++ = wxUnicodePUA + (unsigned char)*opsz; + opsz++; + len++; +#endif + } + } + else if (m_options & MAP_INVALID_UTF8_TO_OCTAL) + { + while (opsz < psz && (!buf || len < n)) + { + if ( buf && len + 3 < n ) + { + unsigned char n = *opsz; + *buf++ = L'\\'; + *buf++ = (wchar_t)( L'0' + n / 0100 ); + *buf++ = (wchar_t)( L'0' + (n % 0100) / 010 ); + *buf++ = (wchar_t)( L'0' + n % 010 ); + } + opsz++; + len += 4; + } + } + else // MAP_INVALID_UTF8_NOT + { + return (size_t)-1; + } } } } @@ -596,6 +743,11 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const return len; } +static inline bool isoctal(wchar_t wch) +{ + return L'0' <= wch && wch <= L'7'; +} + size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const { size_t len = 0; @@ -610,36 +762,67 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const #else cc=(*psz++) & 0x7fffffff; #endif - unsigned cnt; - for (cnt = 0; cc > utf8_max[cnt]; cnt++) {} - if (!cnt) + + if ( (m_options & MAP_INVALID_UTF8_TO_PUA) + && cc >= wxUnicodePUA && cc < wxUnicodePUAEnd ) + { + if (buf) + *buf++ = (char)(cc - wxUnicodePUA); + len++; + } + else if ( (m_options & MAP_INVALID_UTF8_TO_OCTAL) + && cc == L'\\' && psz[0] == L'\\' ) { - // plain ASCII char if (buf) - *buf++ = (char) cc; + *buf++ = (char)cc; + psz++; len++; } + else if ( (m_options & MAP_INVALID_UTF8_TO_OCTAL) && + cc == L'\\' && + isoctal(psz[0]) && isoctal(psz[1]) && isoctal(psz[2]) ) + { + if (buf) + { + *buf++ = (char) ((psz[0] - L'0')*0100 + + (psz[1] - L'0')*010 + + (psz[2] - L'0')); + } + psz += 3; + len++; + } else { - len += cnt + 1; - if (buf) + unsigned cnt; + for (cnt = 0; cc > utf8_max[cnt]; cnt++) {} + if (!cnt) { - *buf++ = (char) ((-128 >> cnt) | ((cc >> (cnt * 6)) & (0x3f >> cnt))); - while (cnt--) - *buf++ = (char) (0x80 | ((cc >> (cnt * 6)) & 0x3f)); + // plain ASCII char + if (buf) + *buf++ = (char) cc; + len++; + } + + else + { + len += cnt + 1; + if (buf) + { + *buf++ = (char) ((-128 >> cnt) | ((cc >> (cnt * 6)) & (0x3f >> cnt))); + while (cnt--) + *buf++ = (char) (0x80 | ((cc >> (cnt * 6)) & 0x3f)); + } } } } - if (buf && (lenWC conversion would fail "randomly". + wxMutexLocker lock(wxConstCast(this, wxMBConv_iconv)->m_iconvMutex); +#endif + size_t inbuf = strlen(psz); size_t outbuf = n * SIZEOF_WCHAR_T; size_t res, cres; @@ -1316,6 +1514,11 @@ size_t wxMBConv_iconv::MB2WC(wchar_t *buf, const char *psz, size_t n) const size_t wxMBConv_iconv::WC2MB(char *buf, const wchar_t *psz, size_t n) const { +#if wxUSE_THREADS + // NB: explained in MB2WC + wxMutexLocker lock(wxConstCast(this, wxMBConv_iconv)->m_iconvMutex); +#endif + size_t inbuf = wxWcslen(psz) * SIZEOF_WCHAR_T; size_t outbuf = n; size_t res, cres; @@ -1419,10 +1622,20 @@ public: // and break the library itself, e.g. wxTextInputStream::NextChar() // wouldn't work if reading an incomplete MB char didn't result in an // error + // + // note however that using MB_ERR_INVALID_CHARS with CP_UTF7 results in + // an error (tested under Windows Server 2003) and apparently it is + // done on purpose, i.e. the function accepts any input in this case + // and although I'd prefer to return error on ill-formed output, our + // own wxMBConvUTF7 doesn't detect errors (e.g. lone "+" which is + // explicitly ill-formed according to RFC 2152) neither so we don't + // even have any fallback here... + int flags = m_CodePage == CP_UTF7 ? 0 : MB_ERR_INVALID_CHARS; + const size_t len = ::MultiByteToWideChar ( m_CodePage, // code page - MB_ERR_INVALID_CHARS, // flags: fall on error + flags, // flags: fall on error psz, // input string -1, // its length (NUL-terminated) buf, // output string @@ -1567,14 +1780,10 @@ private: CFStringEncoding wxCFStringEncFromFontEnc(wxFontEncoding encoding) { - CFStringEncoding enc = 0 ; + CFStringEncoding enc = kCFStringEncodingInvalidId ; if ( encoding == wxFONTENCODING_DEFAULT ) { -#if wxUSE_GUI - encoding = wxFont::GetDefaultEncoding() ; -#else - encoding = wxLocale::GetSystemEncoding() ; -#endif + enc = CFStringGetSystemEncoding(); } else switch( encoding) { @@ -1662,7 +1871,6 @@ CFStringEncoding wxCFStringEncFromFontEnc(wxFontEncoding encoding) case wxFONTENCODING_CP950 : enc = kCFStringEncodingDOSChineseTrad; break ; - case wxFONTENCODING_CP1250 : enc = kCFStringEncodingWindowsLatin2; break ; @@ -1687,9 +1895,10 @@ CFStringEncoding wxCFStringEncFromFontEnc(wxFontEncoding encoding) case wxFONTENCODING_CP1257 : enc = kCFStringEncodingWindowsBalticRim; break ; - case wxFONTENCODING_UTF7 : - enc = kCFStringEncodingNonLossyASCII ; - break ; +// This only really encodes to UTF7 (if that) evidently +// case wxFONTENCODING_UTF7 : +// enc = kCFStringEncodingNonLossyASCII ; +// break ; case wxFONTENCODING_UTF8 : enc = kCFStringEncodingUTF8 ; break ; @@ -1826,245 +2035,6 @@ CFStringEncoding wxCFStringEncFromFontEnc(wxFontEncoding encoding) return enc ; } -wxFontEncoding wxFontEncFromCFStringEnc(CFStringEncoding encoding) -{ - wxFontEncoding enc = wxFONTENCODING_DEFAULT ; - - switch( encoding) - { - case kCFStringEncodingISOLatin1 : - enc = wxFONTENCODING_ISO8859_1 ; - break ; - case kCFStringEncodingISOLatin2 : - enc = wxFONTENCODING_ISO8859_2; - break ; - case kCFStringEncodingISOLatin3 : - enc = wxFONTENCODING_ISO8859_3 ; - break ; - case kCFStringEncodingISOLatin4 : - enc = wxFONTENCODING_ISO8859_4; - break ; - case kCFStringEncodingISOLatinCyrillic : - enc = wxFONTENCODING_ISO8859_5; - break ; - case kCFStringEncodingISOLatinArabic : - enc = wxFONTENCODING_ISO8859_6; - break ; - case kCFStringEncodingISOLatinGreek : - enc = wxFONTENCODING_ISO8859_7; - break ; - case kCFStringEncodingISOLatinHebrew : - enc = wxFONTENCODING_ISO8859_8; - break ; - case kCFStringEncodingISOLatin5 : - enc = wxFONTENCODING_ISO8859_9; - break ; - case kCFStringEncodingISOLatin6 : - enc = wxFONTENCODING_ISO8859_10; - break ; - case kCFStringEncodingISOLatin7 : - enc = wxFONTENCODING_ISO8859_13; - break ; - case kCFStringEncodingISOLatin8 : - enc = wxFONTENCODING_ISO8859_14; - break ; - case kCFStringEncodingISOLatin9 : - enc =wxFONTENCODING_ISO8859_15 ; - break ; - - case kCFStringEncodingKOI8_R : - enc = wxFONTENCODING_KOI8; - break ; - -// case : -// enc = wxFONTENCODING_BULGARIAN; -// break ; - - case kCFStringEncodingDOSLatinUS : - enc = wxFONTENCODING_CP437; - break ; - case kCFStringEncodingDOSLatin1 : - enc = wxFONTENCODING_CP850; - break ; - case kCFStringEncodingDOSLatin2 : - enc =wxFONTENCODING_CP852 ; - break ; - case kCFStringEncodingDOSCyrillic : - enc = wxFONTENCODING_CP855; - break ; - case kCFStringEncodingDOSRussian : - enc = wxFONTENCODING_CP866; - break ; - case kCFStringEncodingDOSThai : - enc =wxFONTENCODING_CP874 ; - break ; - case kCFStringEncodingDOSJapanese : - enc = wxFONTENCODING_CP932; - break ; - case kCFStringEncodingDOSChineseSimplif : - enc = wxFONTENCODING_CP936; - break ; - case kCFStringEncodingDOSKorean : - enc = wxFONTENCODING_CP949; - break ; - case kCFStringEncodingDOSChineseTrad : - enc = wxFONTENCODING_CP950; - break ; - - case kCFStringEncodingWindowsLatin2 : - enc = wxFONTENCODING_CP1250; - break ; - case kCFStringEncodingWindowsCyrillic : - enc = wxFONTENCODING_CP1251; - break ; - case kCFStringEncodingWindowsLatin1 : - enc = wxFONTENCODING_CP1252; - break ; - case kCFStringEncodingWindowsGreek : - enc = wxFONTENCODING_CP1253; - break ; - case kCFStringEncodingWindowsLatin5 : - enc = wxFONTENCODING_CP1254; - break ; - case kCFStringEncodingWindowsHebrew : - enc = wxFONTENCODING_CP1255; - break ; - case kCFStringEncodingWindowsArabic : - enc = wxFONTENCODING_CP1256; - break ; - case kCFStringEncodingWindowsBalticRim : - enc =wxFONTENCODING_CP1257 ; - break ; - case kCFStringEncodingEUC_JP : - enc = wxFONTENCODING_EUC_JP; - break ; - case kCFStringEncodingUnicode : - enc = wxFONTENCODING_UTF16; - break; - case kCFStringEncodingMacRoman : - enc = wxFONTENCODING_MACROMAN ; - break ; - case kCFStringEncodingMacJapanese : - enc = wxFONTENCODING_MACJAPANESE ; - break ; - case kCFStringEncodingMacChineseTrad : - enc = wxFONTENCODING_MACCHINESETRAD ; - break ; - case kCFStringEncodingMacKorean : - enc = wxFONTENCODING_MACKOREAN ; - break ; - case kCFStringEncodingMacArabic : - enc =wxFONTENCODING_MACARABIC ; - break ; - case kCFStringEncodingMacHebrew : - enc = wxFONTENCODING_MACHEBREW ; - break ; - case kCFStringEncodingMacGreek : - enc = wxFONTENCODING_MACGREEK ; - break ; - case kCFStringEncodingMacCyrillic : - enc = wxFONTENCODING_MACCYRILLIC ; - break ; - case kCFStringEncodingMacDevanagari : - enc = wxFONTENCODING_MACDEVANAGARI ; - break ; - case kCFStringEncodingMacGurmukhi : - enc = wxFONTENCODING_MACGURMUKHI ; - break ; - case kCFStringEncodingMacGujarati : - enc = wxFONTENCODING_MACGUJARATI ; - break ; - case kCFStringEncodingMacOriya : - enc =wxFONTENCODING_MACORIYA ; - break ; - case kCFStringEncodingMacBengali : - enc =wxFONTENCODING_MACBENGALI ; - break ; - case kCFStringEncodingMacTamil : - enc = wxFONTENCODING_MACTAMIL ; - break ; - case kCFStringEncodingMacTelugu : - enc = wxFONTENCODING_MACTELUGU ; - break ; - case kCFStringEncodingMacKannada : - enc = wxFONTENCODING_MACKANNADA ; - break ; - case kCFStringEncodingMacMalayalam : - enc = wxFONTENCODING_MACMALAJALAM ; - break ; - case kCFStringEncodingMacSinhalese : - enc = wxFONTENCODING_MACSINHALESE ; - break ; - case kCFStringEncodingMacBurmese : - enc = wxFONTENCODING_MACBURMESE ; - break ; - case kCFStringEncodingMacKhmer : - enc = wxFONTENCODING_MACKHMER ; - break ; - case kCFStringEncodingMacThai : - enc = wxFONTENCODING_MACTHAI ; - break ; - case kCFStringEncodingMacLaotian : - enc = wxFONTENCODING_MACLAOTIAN ; - break ; - case kCFStringEncodingMacGeorgian : - enc = wxFONTENCODING_MACGEORGIAN ; - break ; - case kCFStringEncodingMacArmenian : - enc = wxFONTENCODING_MACARMENIAN ; - break ; - case kCFStringEncodingMacChineseSimp : - enc = wxFONTENCODING_MACCHINESESIMP ; - break ; - case kCFStringEncodingMacTibetan : - enc = wxFONTENCODING_MACTIBETAN ; - break ; - case kCFStringEncodingMacMongolian : - enc = wxFONTENCODING_MACMONGOLIAN ; - break ; - case kCFStringEncodingMacEthiopic : - enc = wxFONTENCODING_MACETHIOPIC ; - break ; - case kCFStringEncodingMacCentralEurRoman: - enc = wxFONTENCODING_MACCENTRALEUR ; - break ; - case kCFStringEncodingMacVietnamese: - enc = wxFONTENCODING_MACVIATNAMESE ; - break ; - case kCFStringEncodingMacExtArabic : - enc = wxFONTENCODING_MACARABICEXT ; - break ; - case kCFStringEncodingMacSymbol : - enc = wxFONTENCODING_MACSYMBOL ; - break ; - case kCFStringEncodingMacDingbats : - enc = wxFONTENCODING_MACDINGBATS ; - break ; - case kCFStringEncodingMacTurkish : - enc = wxFONTENCODING_MACTURKISH ; - break ; - case kCFStringEncodingMacCroatian : - enc = wxFONTENCODING_MACCROATIAN ; - break ; - case kCFStringEncodingMacIcelandic : - enc = wxFONTENCODING_MACICELANDIC ; - break ; - case kCFStringEncodingMacRomanian : - enc = wxFONTENCODING_MACROMANIAN ; - break ; - case kCFStringEncodingMacCeltic : - enc = wxFONTENCODING_MACCELTIC ; - break ; - case kCFStringEncodingMacGaelic : - enc = wxFONTENCODING_MACGAELIC ; - break ; -// case kCFStringEncodingMacKeyboardGlyphs : -// enc = wxFONTENCODING_MACKEYBOARD ; -// break ; - } ; - return enc ; -} - class wxMBConv_cocoa : public wxMBConv { public: @@ -2073,10 +2043,12 @@ public: Init(CFStringGetSystemEncoding()) ; } +#if wxUSE_FONTMAP wxMBConv_cocoa(const wxChar* name) { - Init( wxCFStringEncFromFontEnc(wxFontMapper::Get()->CharsetToEncoding(name, false) ) ) ; + Init( wxCFStringEncFromFontEnc(wxFontMapperBase::Get()->CharsetToEncoding(name, false) ) ) ; } +#endif wxMBConv_cocoa(wxFontEncoding encoding) { @@ -2089,160 +2061,118 @@ public: void Init( CFStringEncoding encoding) { - m_char_encoding = encoding ; - m_unicode_encoding = kCFStringEncodingUnicode; + m_encoding = encoding ; } size_t MB2WC(wchar_t * szOut, const char * szUnConv, size_t nOutSize) const { wxASSERT(szUnConv); - size_t nBufSize = strlen(szUnConv) + 1; - size_t nRealOutSize; - - UniChar* szUniCharBuffer = (UniChar*) szOut; - wchar_t* szConvBuffer = szOut; - - if (szConvBuffer == NULL && nOutSize != 0) - { - szConvBuffer = new wchar_t[nOutSize] ; - } - -#if SIZEOF_WCHAR_T == 4 - szUniCharBuffer = new UniChar[nOutSize]; -#endif - - CFDataRef theData = CFDataCreateWithBytesNoCopy ( - NULL, //allocator - (const UInt8*)szUnConv, - nBufSize - 1, - NULL //deallocator - ); - - wxASSERT(theData); - - CFStringRef theString = CFStringCreateFromExternalRepresentation ( - NULL, - theData, - m_char_encoding + CFStringRef theString = CFStringCreateWithBytes ( + NULL, //the allocator + (const UInt8*)szUnConv, + strlen(szUnConv), + m_encoding, + false //no BOM/external representation ); wxASSERT(theString); - if (nOutSize == 0) + size_t nOutLength = CFStringGetLength(theString); + + if (szOut == NULL) { - nRealOutSize = CFStringGetLength(theString) + 1; CFRelease(theString); - return nRealOutSize - 1; + return nOutLength; } - CFRange theRange = { 0, CFStringGetLength(theString) }; - - CFStringGetCharacters(theString, theRange, szUniCharBuffer); + CFRange theRange = { 0, nOutSize }; +#if SIZEOF_WCHAR_T == 4 + UniChar* szUniCharBuffer = new UniChar[nOutSize]; +#endif - nRealOutSize = (CFStringGetLength(theString) + 1); + CFStringGetCharacters(theString, theRange, szUniCharBuffer); CFRelease(theString); - szUniCharBuffer[nRealOutSize-1] = '\0' ; + szUniCharBuffer[nOutLength] = '\0' ; #if SIZEOF_WCHAR_T == 4 wxMBConvUTF16 converter ; - converter.MB2WC(szConvBuffer , (const char*)szUniCharBuffer , nRealOutSize ) ; + converter.MB2WC(szOut, (const char*)szUniCharBuffer , nOutSize ) ; delete[] szUniCharBuffer; #endif - if ( szOut == NULL ) - delete [] szConvBuffer; - return nRealOutSize ; + return nOutLength; } size_t WC2MB(char *szOut, const wchar_t *szUnConv, size_t nOutSize) const { - size_t nBufSize = wxWcslen(szUnConv) + 1; + wxASSERT(szUnConv); + size_t nRealOutSize; - char* szBuffer = szOut; + size_t nBufSize = wxWcslen(szUnConv); UniChar* szUniBuffer = (UniChar*) szUnConv; - if (szOut == NULL) - { - // worst case - nRealOutSize = ((nBufSize - 1) * 8) +1 ; - szBuffer = new char[ nRealOutSize ] ; - } - else - nRealOutSize = nOutSize; - #if SIZEOF_WCHAR_T == 4 - wxMBConvUTF16BE converter ; + wxMBConvUTF16 converter ; nBufSize = converter.WC2MB( NULL , szUnConv , 0 ); szUniBuffer = new UniChar[ (nBufSize / sizeof(UniChar)) + 1] ; converter.WC2MB( (char*) szUniBuffer , szUnConv, nBufSize + sizeof(UniChar)) ; nBufSize /= sizeof(UniChar); - ++nBufSize; #endif CFStringRef theString = CFStringCreateWithCharactersNoCopy( NULL, //allocator szUniBuffer, nBufSize, - NULL //deallocator + kCFAllocatorNull //deallocator - we want to deallocate it ourselves ); wxASSERT(theString); //Note that CER puts a BOM when converting to unicode - //so we may want to check and use getchars instead in that case - CFDataRef theData = CFStringCreateExternalRepresentation( - NULL, //allocator - theString, - m_char_encoding, - 0 //what to put in characters that can't be converted - - //0 tells CFString to return NULL if it meets such a character - ); - - if(!theData) - return (size_t)-1; - - CFRelease(theString); - - nRealOutSize = CFDataGetLength(theData); - - if ( szOut == NULL ) - delete[] szBuffer; + //so we check and use getchars instead in that case + if (m_encoding == kCFStringEncodingUnicode) + { + if (szOut != NULL) + CFStringGetCharacters(theString, CFRangeMake(0, nOutSize - 1), (UniChar*) szOut); - if(nOutSize == 0) + nRealOutSize = CFStringGetLength(theString) + 1; + } + else { -//TODO: This gets flagged as a non-malloced address by the debugger... -//#if SIZEOF_WCHAR_T == 4 -// delete[] szUniBuffer; -//#endif - CFRelease(theData); - return nRealOutSize - 1; + CFStringGetBytes( + theString, + CFRangeMake(0, CFStringGetLength(theString)), + m_encoding, + 0, //what to put in characters that can't be converted - + //0 tells CFString to return NULL if it meets such a character + false, //not an external representation + (UInt8*) szOut, + nOutSize, + (CFIndex*) &nRealOutSize + ); } - CFRange theRange = {0, CFDataGetLength(theData) }; - CFDataGetBytes(theData, theRange, (UInt8*) szBuffer); + CFRelease(theString); - CFRelease(theData); +#if SIZEOF_WCHAR_T == 4 + delete[] szUniBuffer; +#endif -//TODO: This gets flagged as a non-malloced address by the debugger... -//#if SIZEOF_WCHAR_T == 4 -// delete[] szUniBuffer; -//#endif return nRealOutSize - 1; } bool IsOk() const { - //TODO: check for invalid en/de/coding - return true; + return m_encoding != kCFStringEncodingInvalidId && + CFStringIsEncodingAvailable(m_encoding); } private: - CFStringEncoding m_char_encoding ; - CFStringEncoding m_unicode_encoding ; + CFStringEncoding m_encoding ; }; #endif // defined(__WXCOCOA__) @@ -2261,10 +2191,12 @@ public: Init(CFStringGetSystemEncoding()) ; } +#if wxUSE_FONTMAP wxMBConv_mac(const wxChar* name) { - Init( wxMacGetSystemEncFromFontEnc(wxFontMapper::Get()->CharsetToEncoding(name, false) ) ) ; + Init( wxMacGetSystemEncFromFontEnc(wxFontMapperBase::Get()->CharsetToEncoding(name, false) ) ) ; } +#endif wxMBConv_mac(wxFontEncoding encoding) { @@ -2304,7 +2236,8 @@ public: if (buf == NULL) { - n = byteInLen ; + //apple specs say at least 32 + n = wxMax( 32 , byteInLen ) ; tbuf = (wchar_t*) malloc( n * SIZEOF_WCHAR_T) ; } ByteCount byteBufferLen = n * sizeof( UniChar ) ; @@ -2319,7 +2252,7 @@ public: // we have to terminate here, because n might be larger for the trailing zero, and if UniChar // is not properly terminated we get random characters at the end ubuf[byteOutLen / sizeof( UniChar ) ] = 0 ; - wxMBConvUTF16BE converter ; + wxMBConvUTF16 converter ; res = converter.MB2WC( (buf ? buf : tbuf) , (const char*)ubuf , n ) ; free( ubuf ) ; #else @@ -2334,13 +2267,6 @@ public: return res ; } - //NB: This is _broken_ - in invalid conversions, instead of returning -1 - //like it should, it (sometimes?) converts invalid characters of the encoding to a question - //mark character '?'. - // - //We need to do the msw double-pass check for the question marks as Vadim - //lines out above (RN: I don't recall this happening in the core foundation version, - //but it might do it there also, ick) size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const { OSStatus status = noErr ; @@ -2351,15 +2277,15 @@ public: if (buf == NULL) { - // worst case - n = ((byteInLen / SIZEOF_WCHAR_T) * 8) + SIZEOF_WCHAR_T; + //apple specs say at least 32 + n = wxMax( 32 , ((byteInLen / SIZEOF_WCHAR_T) * 8) + SIZEOF_WCHAR_T ); tbuf = (char*) malloc( n ) ; } ByteCount byteBufferLen = n ; UniChar* ubuf = NULL ; #if SIZEOF_WCHAR_T == 4 - wxMBConvUTF16BE converter ; + wxMBConvUTF16 converter ; size_t unicharlen = converter.WC2MB( NULL , psz , 0 ) ; byteInLen = unicharlen ; ubuf = (UniChar*) malloc( byteInLen + 2 ) ; @@ -2377,8 +2303,23 @@ public: size_t res = byteOutLen ; if ( buf && res < n) + { buf[res] = 0; + //we need to double-trip to verify it didn't insert any ? in place + //of bogus characters + wxWCharBuffer wcBuf(n); + size_t pszlen = wxWcslen(psz); + if ( MB2WC(wcBuf.data(), buf, n) == (size_t)-1 || + wxWcslen(wcBuf) != pszlen || + memcmp(wcBuf, psz, pszlen * sizeof(wchar_t)) != 0 ) + { + // we didn't obtain the same thing we started from, hence + // the conversion was lossy and we consider that it failed + return (size_t)-1; + } + } + return res ; } @@ -2416,7 +2357,7 @@ public: wxMBConv_wxwin(const wxChar* name) { if (name) - m_enc = wxFontMapper::Get()->CharsetToEncoding(name, false); + m_enc = wxFontMapperBase::Get()->CharsetToEncoding(name, false); else m_enc = wxFONTENCODING_SYSTEM; @@ -2434,7 +2375,10 @@ public: { size_t inbuf = strlen(psz); if (buf) - m2w.Convert(psz,buf); + { + if (!m2w.Convert(psz,buf)) + return (size_t)-1; + } return inbuf; } @@ -2442,7 +2386,10 @@ public: { const size_t inbuf = wxWcslen(psz); if (buf) - w2m.Convert(psz,buf); + { + if (!w2m.Convert(psz,buf)) + return (size_t)-1; + } return inbuf; } @@ -2569,7 +2516,7 @@ wxMBConv *wxCSConv::DoCreate() const #if wxUSE_FONTMAP if ( name.empty() ) - name = wxFontMapper::Get()->GetEncodingName(m_encoding); + name = wxFontMapperBase::Get()->GetEncodingName(m_encoding); #endif // wxUSE_FONTMAP wxMBConv_iconv *conv = new wxMBConv_iconv(name); @@ -2596,11 +2543,17 @@ wxMBConv *wxCSConv::DoCreate() const #endif // wxHAVE_WIN32_MB2WC #if defined(__WXMAC__) { - if ( m_name || ( m_encoding < wxFONTENCODING_UTF16BE ) ) + // leave UTF16 and UTF32 to the built-ins of wx + if ( m_name || ( m_encoding < wxFONTENCODING_UTF16BE || + ( m_encoding >= wxFONTENCODING_MACMIN && m_encoding <= wxFONTENCODING_MACMAX ) ) ) { +#if wxUSE_FONTMAP wxMBConv_mac *conv = m_name ? new wxMBConv_mac(m_name) : new wxMBConv_mac(m_encoding); +#else + wxMBConv_mac *conv = new wxMBConv_mac(m_encoding); +#endif if ( conv->IsOk() ) return conv; @@ -2613,8 +2566,12 @@ wxMBConv *wxCSConv::DoCreate() const if ( m_name || ( m_encoding <= wxFONTENCODING_UTF16 ) ) { +#if wxUSE_FONTMAP wxMBConv_cocoa *conv = m_name ? new wxMBConv_cocoa(m_name) : new wxMBConv_cocoa(m_encoding); +#else + wxMBConv_cocoa *conv = new wxMBConv_cocoa(m_encoding); +#endif if ( conv->IsOk() ) return conv; @@ -2630,7 +2587,7 @@ wxMBConv *wxCSConv::DoCreate() const // 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); + enc = wxFontMapperBase::Get()->CharsetToEncoding(m_name, false); } #endif // wxUSE_FONTMAP @@ -2686,7 +2643,7 @@ wxMBConv *wxCSConv::DoCreate() const m_name ? m_name : #if wxUSE_FONTMAP - wxFontMapper::GetEncodingDescription(m_encoding).c_str() + wxFontMapperBase::GetEncodingDescription(m_encoding).c_str() #else // !wxUSE_FONTMAP wxString::Format(_("encoding %s"), m_encoding).c_str() #endif // wxUSE_FONTMAP/!wxUSE_FONTMAP @@ -2783,13 +2740,19 @@ 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; +WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName = & +#ifdef __WXOSX__ + wxConvUTF8Obj; +#else + wxConvLibcObj; +#endif + #else // !wxUSE_WCHAR_T