X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/faa60a4f0b456f6eed431c4dd224c1ec04252d99..c2331d9b50dffbd3f2c0caf06b514fd4c96b3b94:/src/common/strconv.cpp?ds=sidebyside diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 05aa1faa3e..609f44b9d1 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -56,16 +56,9 @@ #include "wx/encconv.h" #include "wx/fontmap.h" -#ifdef __WXMAC__ -#ifndef __DARWIN__ -#include -#include -#include -#endif - -// includes Mac headers -#include "wx/mac/private.h" -#endif +#ifdef __DARWIN__ +#include "wx/mac/corefoundation/private/strconv_cf.h" +#endif //def __DARWIN__ #define TRACE_STRCONV _T("strconv") @@ -361,14 +354,14 @@ const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const if ( psz ) { // calculate the length of the buffer needed first - const size_t nLen = MB2WC(NULL, psz, 0); + const size_t nLen = ToWChar(NULL, 0, psz); if ( nLen != wxCONV_FAILED ) { // now do the actual conversion - wxWCharBuffer buf(nLen /* +1 added implicitly */); + wxWCharBuffer buf(nLen - 1 /* +1 added implicitly */); // +1 for the trailing NULL - if ( MB2WC(buf.data(), psz, nLen + 1) != wxCONV_FAILED ) + if ( ToWChar(buf.data(), nLen, psz) != wxCONV_FAILED ) return buf; } } @@ -380,14 +373,11 @@ const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *pwz) const { if ( pwz ) { - const size_t nLen = WC2MB(NULL, pwz, 0); + const size_t nLen = FromWChar(NULL, 0, pwz); if ( nLen != wxCONV_FAILED ) { - // extra space for trailing NUL(s) - static const size_t extraLen = GetMaxMBNulLen(); - - wxCharBuffer buf(nLen + extraLen - 1); - if ( WC2MB(buf.data(), pwz, nLen + extraLen) != wxCONV_FAILED ) + wxCharBuffer buf(nLen - 1); + if ( FromWChar(buf.data(), nLen, pwz) != wxCONV_FAILED ) return buf; } } @@ -713,7 +703,7 @@ size_t wxMBConvUTF7::WC2MB(char *buf, const wchar_t *psz, size_t n) const // UTF-8 // ---------------------------------------------------------------------------- -static wxUint32 utf8_max[]= +static const wxUint32 utf8_max[]= { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff, 0xffffffff }; // boundaries of the private use area we use to (temporarily) remap invalid @@ -721,8 +711,283 @@ static wxUint32 utf8_max[]= const wxUint32 wxUnicodePUA = 0x100000; const wxUint32 wxUnicodePUAEnd = wxUnicodePUA + 256; +// this table gives the length of the UTF-8 encoding from its first character: +const unsigned char tableUtf8Lengths[256] = { + // single-byte sequences (ASCII): + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 00..0F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 10..1F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 20..2F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 30..3F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 40..4F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 50..5F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 60..6F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 70..7F + + // these are invalid: + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 80..8F + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 90..9F + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // A0..AF + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // B0..BF + 0, 0, // C0,C1 + + // two-byte sequences: + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // C2..CF + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // D0..DF + + // three-byte sequences: + 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, // E0..EF + + // four-byte sequences: + 4, 4, 4, 4, 4, // F0..F4 + + // these are invalid again (5- or 6-byte + // sequences and sequences for code points + // above U+10FFFF, as restricted by RFC 3629): + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 // F5..FF +}; + +size_t +wxMBConvStrictUTF8::ToWChar(wchar_t *dst, size_t dstLen, + const char *src, size_t srcLen) const +{ + wchar_t *out = dstLen ? dst : NULL; + size_t written = 0; + + if ( srcLen == wxNO_LEN ) + srcLen = strlen(src) + 1; + + for ( const char *p = src; ; p++ ) + { + if ( !(srcLen == wxNO_LEN ? *p : srcLen) ) + { + // all done successfully, just add the trailing NULL if we are not + // using explicit length + if ( srcLen == wxNO_LEN ) + { + if ( out ) + { + if ( !dstLen ) + break; + + *out = L'\0'; + } + + written++; + } + + return written; + } + + if ( out && !dstLen-- ) + break; + + wxUint32 code; + unsigned char c = *p; + + if ( c < 0x80 ) + { + if ( srcLen == 0 ) // the test works for wxNO_LEN too + break; + + if ( srcLen != wxNO_LEN ) + srcLen--; + + code = c; + } + else + { + unsigned len = tableUtf8Lengths[c]; + if ( !len ) + break; + + if ( srcLen < len ) // the test works for wxNO_LEN too + break; + + if ( srcLen != wxNO_LEN ) + srcLen -= len; + + // Char. number range | UTF-8 octet sequence + // (hexadecimal) | (binary) + // ----------------------+---------------------------------------- + // 0000 0000 - 0000 007F | 0xxxxxxx + // 0000 0080 - 0000 07FF | 110xxxxx 10xxxxxx + // 0000 0800 - 0000 FFFF | 1110xxxx 10xxxxxx 10xxxxxx + // 0001 0000 - 0010 FFFF | 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx + // + // Code point value is stored in bits marked with 'x', + // lowest-order bit of the value on the right side in the diagram + // above. (from RFC 3629) + + // mask to extract lead byte's value ('x' bits above), by sequence + // length: + static const unsigned char leadValueMask[] = { 0x7F, 0x1F, 0x0F, 0x07 }; + + // mask and value of lead byte's most significant bits, by length: + static const unsigned char leadMarkerMask[] = { 0x80, 0xE0, 0xF0, 0xF8 }; + static const unsigned char leadMarkerVal[] = { 0x00, 0xC0, 0xE0, 0xF0 }; + + len--; // it's more convenient to work with 0-based length here + + // extract the lead byte's value bits: + if ( (c & leadMarkerMask[len]) != leadMarkerVal[len] ) + break; + + code = c & leadValueMask[len]; + + // all remaining bytes, if any, are handled in the same way + // regardless of sequence's length: + for ( ; len; --len ) + { + c = *++p; + if ( (c & 0xC0) != 0x80 ) + return wxCONV_FAILED; + + code <<= 6; + code |= c & 0x3F; + } + } + +#ifdef WC_UTF16 + // cast is ok because wchar_t == wxUint16 if WC_UTF16 + if ( encode_utf16(code, (wxUint16 *)out) == 2 ) + { + if ( out ) + out++; + written++; + } +#else // !WC_UTF16 + if ( out ) + *out = code; +#endif // WC_UTF16/!WC_UTF16 + + if ( out ) + out++; + + written++; + } + + return wxCONV_FAILED; +} + +size_t +wxMBConvStrictUTF8::FromWChar(char *dst, size_t dstLen, + const wchar_t *src, size_t srcLen) const +{ + char *out = dstLen ? dst : NULL; + size_t written = 0; + + for ( const wchar_t *wp = src; ; wp++ ) + { + if ( !(srcLen == wxNO_LEN ? *wp : srcLen--) ) + { + // all done successfully, just add the trailing NULL if we are not + // using explicit length + if ( srcLen == wxNO_LEN ) + { + if ( out ) + { + if ( !dstLen ) + break; + + *out = '\0'; + } + + written++; + } + + return written; + } + + + wxUint32 code; +#ifdef WC_UTF16 + // cast is ok for WC_UTF16 + if ( decode_utf16((const wxUint16 *)wp, code) == 2 ) + { + // skip the next char too as we decoded a surrogate + wp++; + } +#else // wchar_t is UTF-32 + code = *wp & 0x7fffffff; +#endif + + unsigned len; + if ( code <= 0x7F ) + { + len = 1; + if ( out ) + { + if ( dstLen < len ) + break; + + out[0] = (char)code; + } + } + else if ( code <= 0x07FF ) + { + len = 2; + if ( out ) + { + if ( dstLen < len ) + break; + + // NB: this line takes 6 least significant bits, encodes them as + // 10xxxxxx and discards them so that the next byte can be encoded: + out[1] = 0x80 | (code & 0x3F); code >>= 6; + out[0] = 0xC0 | code; + } + } + else if ( code < 0xFFFF ) + { + len = 3; + if ( out ) + { + if ( dstLen < len ) + break; + + out[2] = 0x80 | (code & 0x3F); code >>= 6; + out[1] = 0x80 | (code & 0x3F); code >>= 6; + out[0] = 0xE0 | code; + } + } + else if ( code <= 0x10FFFF ) + { + len = 4; + if ( out ) + { + if ( dstLen < len ) + break; + + out[3] = 0x80 | (code & 0x3F); code >>= 6; + out[2] = 0x80 | (code & 0x3F); code >>= 6; + out[1] = 0x80 | (code & 0x3F); code >>= 6; + out[0] = 0xF0 | code; + } + } + else + { + wxFAIL_MSG( _T("trying to encode undefined Unicode character") ); + break; + } + + if ( out ) + { + out += len; + dstLen -= len; + } + + written += len; + } + + // we only get here if an error occurs during decoding + return wxCONV_FAILED; +} + size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const { + if ( m_options == MAP_INVALID_UTF8_NOT ) + return wxMBConvStrictUTF8::MB2WC(buf, psz, n); + size_t len = 0; while (*psz && ((!buf) || (len < n))) @@ -792,7 +1057,7 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const else { #ifdef WC_UTF16 - // cast is ok because wchar_t == wxUuint16 if WC_UTF16 + // cast is ok because wchar_t == wxUint16 if WC_UTF16 size_t pa = encode_utf16(res, (wxUint16 *)buf); if (pa == wxCONV_FAILED) { @@ -872,6 +1137,9 @@ static inline bool isoctal(wchar_t wch) size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const { + if ( m_options == MAP_INVALID_UTF8_NOT ) + return wxMBConvStrictUTF8::WC2MB(buf, psz, n); + size_t len = 0; while (*psz && ((!buf) || (len < n))) @@ -1746,8 +2014,8 @@ wxMBConv_iconv::wxMBConv_iconv(const char *name) wxLogTrace(TRACE_STRCONV, wxT("iconv wchar_t charset is \"%s\"%s"), - ms_wcCharsetName.empty() ? _T("") - : ms_wcCharsetName.c_str(), + ms_wcCharsetName.empty() ? wxString("") + : ms_wcCharsetName, ms_wcNeedsSwap ? _T(" (needs swap)") : _T("")); } @@ -1956,7 +2224,7 @@ size_t wxMBConv_iconv::GetMBNulLen() const wxMutexLocker lock(self->m_iconvMutex); #endif - wchar_t *wnul = L""; + const wchar_t *wnul = L""; char buf[8]; // should be enough for NUL in any encoding size_t inLen = sizeof(wchar_t), outLen = WXSIZEOF(buf); @@ -2296,794 +2564,6 @@ private: #endif // wxHAVE_WIN32_MB2WC -// ============================================================================ -// Cocoa conversion classes -// ============================================================================ - -#if defined(__WXCOCOA__) - -// RN: There is no UTF-32 support in either Core Foundation or Cocoa. -// Strangely enough, internally Core Foundation uses -// UTF-32 internally quite a bit - its just not public (yet). - -#include -#include - -CFStringEncoding wxCFStringEncFromFontEnc(wxFontEncoding encoding) -{ - CFStringEncoding enc = kCFStringEncodingInvalidId ; - - switch (encoding) - { - case wxFONTENCODING_DEFAULT : - enc = CFStringGetSystemEncoding(); - break ; - - case wxFONTENCODING_ISO8859_1 : - enc = kCFStringEncodingISOLatin1 ; - break ; - case wxFONTENCODING_ISO8859_2 : - enc = kCFStringEncodingISOLatin2; - break ; - case wxFONTENCODING_ISO8859_3 : - enc = kCFStringEncodingISOLatin3 ; - break ; - case wxFONTENCODING_ISO8859_4 : - enc = kCFStringEncodingISOLatin4; - break ; - case wxFONTENCODING_ISO8859_5 : - enc = kCFStringEncodingISOLatinCyrillic; - break ; - case wxFONTENCODING_ISO8859_6 : - enc = kCFStringEncodingISOLatinArabic; - break ; - case wxFONTENCODING_ISO8859_7 : - enc = kCFStringEncodingISOLatinGreek; - break ; - case wxFONTENCODING_ISO8859_8 : - enc = kCFStringEncodingISOLatinHebrew; - break ; - case wxFONTENCODING_ISO8859_9 : - enc = kCFStringEncodingISOLatin5; - break ; - case wxFONTENCODING_ISO8859_10 : - enc = kCFStringEncodingISOLatin6; - break ; - case wxFONTENCODING_ISO8859_11 : - enc = kCFStringEncodingISOLatinThai; - break ; - case wxFONTENCODING_ISO8859_13 : - enc = kCFStringEncodingISOLatin7; - break ; - case wxFONTENCODING_ISO8859_14 : - enc = kCFStringEncodingISOLatin8; - break ; - case wxFONTENCODING_ISO8859_15 : - enc = kCFStringEncodingISOLatin9; - break ; - - case wxFONTENCODING_KOI8 : - enc = kCFStringEncodingKOI8_R; - break ; - case wxFONTENCODING_ALTERNATIVE : // MS-DOS CP866 - enc = kCFStringEncodingDOSRussian; - break ; - -// case wxFONTENCODING_BULGARIAN : -// enc = ; -// break ; - - case wxFONTENCODING_CP437 : - enc = kCFStringEncodingDOSLatinUS ; - break ; - case wxFONTENCODING_CP850 : - enc = kCFStringEncodingDOSLatin1; - break ; - case wxFONTENCODING_CP852 : - enc = kCFStringEncodingDOSLatin2; - break ; - case wxFONTENCODING_CP855 : - enc = kCFStringEncodingDOSCyrillic; - break ; - case wxFONTENCODING_CP866 : - enc = kCFStringEncodingDOSRussian ; - break ; - case wxFONTENCODING_CP874 : - enc = kCFStringEncodingDOSThai; - break ; - case wxFONTENCODING_CP932 : - enc = kCFStringEncodingDOSJapanese; - break ; - case wxFONTENCODING_CP936 : - enc = kCFStringEncodingDOSChineseSimplif ; - break ; - case wxFONTENCODING_CP949 : - enc = kCFStringEncodingDOSKorean; - break ; - case wxFONTENCODING_CP950 : - enc = kCFStringEncodingDOSChineseTrad; - break ; - case wxFONTENCODING_CP1250 : - enc = kCFStringEncodingWindowsLatin2; - break ; - case wxFONTENCODING_CP1251 : - enc = kCFStringEncodingWindowsCyrillic ; - break ; - case wxFONTENCODING_CP1252 : - enc = kCFStringEncodingWindowsLatin1 ; - break ; - case wxFONTENCODING_CP1253 : - enc = kCFStringEncodingWindowsGreek; - break ; - case wxFONTENCODING_CP1254 : - enc = kCFStringEncodingWindowsLatin5; - break ; - case wxFONTENCODING_CP1255 : - enc = kCFStringEncodingWindowsHebrew ; - break ; - case wxFONTENCODING_CP1256 : - enc = kCFStringEncodingWindowsArabic ; - break ; - case wxFONTENCODING_CP1257 : - enc = kCFStringEncodingWindowsBalticRim; - break ; -// This only really encodes to UTF7 (if that) evidently -// case wxFONTENCODING_UTF7 : -// enc = kCFStringEncodingNonLossyASCII ; -// break ; - case wxFONTENCODING_UTF8 : - enc = kCFStringEncodingUTF8 ; - break ; - case wxFONTENCODING_EUC_JP : - enc = kCFStringEncodingEUC_JP; - break ; - case wxFONTENCODING_UTF16 : - enc = kCFStringEncodingUnicode ; - break ; - case wxFONTENCODING_MACROMAN : - enc = kCFStringEncodingMacRoman ; - break ; - case wxFONTENCODING_MACJAPANESE : - enc = kCFStringEncodingMacJapanese ; - break ; - case wxFONTENCODING_MACCHINESETRAD : - enc = kCFStringEncodingMacChineseTrad ; - break ; - case wxFONTENCODING_MACKOREAN : - enc = kCFStringEncodingMacKorean ; - break ; - case wxFONTENCODING_MACARABIC : - enc = kCFStringEncodingMacArabic ; - break ; - case wxFONTENCODING_MACHEBREW : - enc = kCFStringEncodingMacHebrew ; - break ; - case wxFONTENCODING_MACGREEK : - enc = kCFStringEncodingMacGreek ; - break ; - case wxFONTENCODING_MACCYRILLIC : - enc = kCFStringEncodingMacCyrillic ; - break ; - case wxFONTENCODING_MACDEVANAGARI : - enc = kCFStringEncodingMacDevanagari ; - break ; - case wxFONTENCODING_MACGURMUKHI : - enc = kCFStringEncodingMacGurmukhi ; - break ; - case wxFONTENCODING_MACGUJARATI : - enc = kCFStringEncodingMacGujarati ; - break ; - case wxFONTENCODING_MACORIYA : - enc = kCFStringEncodingMacOriya ; - break ; - case wxFONTENCODING_MACBENGALI : - enc = kCFStringEncodingMacBengali ; - break ; - case wxFONTENCODING_MACTAMIL : - enc = kCFStringEncodingMacTamil ; - break ; - case wxFONTENCODING_MACTELUGU : - enc = kCFStringEncodingMacTelugu ; - break ; - case wxFONTENCODING_MACKANNADA : - enc = kCFStringEncodingMacKannada ; - break ; - case wxFONTENCODING_MACMALAJALAM : - enc = kCFStringEncodingMacMalayalam ; - break ; - case wxFONTENCODING_MACSINHALESE : - enc = kCFStringEncodingMacSinhalese ; - break ; - case wxFONTENCODING_MACBURMESE : - enc = kCFStringEncodingMacBurmese ; - break ; - case wxFONTENCODING_MACKHMER : - enc = kCFStringEncodingMacKhmer ; - break ; - case wxFONTENCODING_MACTHAI : - enc = kCFStringEncodingMacThai ; - break ; - case wxFONTENCODING_MACLAOTIAN : - enc = kCFStringEncodingMacLaotian ; - break ; - case wxFONTENCODING_MACGEORGIAN : - enc = kCFStringEncodingMacGeorgian ; - break ; - case wxFONTENCODING_MACARMENIAN : - enc = kCFStringEncodingMacArmenian ; - break ; - case wxFONTENCODING_MACCHINESESIMP : - enc = kCFStringEncodingMacChineseSimp ; - break ; - case wxFONTENCODING_MACTIBETAN : - enc = kCFStringEncodingMacTibetan ; - break ; - case wxFONTENCODING_MACMONGOLIAN : - enc = kCFStringEncodingMacMongolian ; - break ; - case wxFONTENCODING_MACETHIOPIC : - enc = kCFStringEncodingMacEthiopic ; - break ; - case wxFONTENCODING_MACCENTRALEUR : - enc = kCFStringEncodingMacCentralEurRoman ; - break ; - case wxFONTENCODING_MACVIATNAMESE : - enc = kCFStringEncodingMacVietnamese ; - break ; - case wxFONTENCODING_MACARABICEXT : - enc = kCFStringEncodingMacExtArabic ; - break ; - case wxFONTENCODING_MACSYMBOL : - enc = kCFStringEncodingMacSymbol ; - break ; - case wxFONTENCODING_MACDINGBATS : - enc = kCFStringEncodingMacDingbats ; - break ; - case wxFONTENCODING_MACTURKISH : - enc = kCFStringEncodingMacTurkish ; - break ; - case wxFONTENCODING_MACCROATIAN : - enc = kCFStringEncodingMacCroatian ; - break ; - case wxFONTENCODING_MACICELANDIC : - enc = kCFStringEncodingMacIcelandic ; - break ; - case wxFONTENCODING_MACROMANIAN : - enc = kCFStringEncodingMacRomanian ; - break ; - case wxFONTENCODING_MACCELTIC : - enc = kCFStringEncodingMacCeltic ; - break ; - case wxFONTENCODING_MACGAELIC : - enc = kCFStringEncodingMacGaelic ; - break ; -// case wxFONTENCODING_MACKEYBOARD : -// enc = kCFStringEncodingMacKeyboardGlyphs ; -// break ; - - default : - // because gcc is picky - break ; - } - - return enc ; -} - -class wxMBConv_cocoa : public wxMBConv -{ -public: - wxMBConv_cocoa() - { - Init(CFStringGetSystemEncoding()) ; - } - - wxMBConv_cocoa(const wxMBConv_cocoa& conv) - { - m_encoding = conv.m_encoding; - } - -#if wxUSE_FONTMAP - wxMBConv_cocoa(const wxChar* name) - { - Init( wxCFStringEncFromFontEnc(wxFontMapperBase::Get()->CharsetToEncoding(name, false) ) ) ; - } -#endif - - wxMBConv_cocoa(wxFontEncoding encoding) - { - Init( wxCFStringEncFromFontEnc(encoding) ); - } - - virtual ~wxMBConv_cocoa() - { - } - - void Init( CFStringEncoding encoding) - { - m_encoding = encoding ; - } - - size_t MB2WC(wchar_t * szOut, const char * szUnConv, size_t nOutSize) const - { - wxASSERT(szUnConv); - - CFStringRef theString = CFStringCreateWithBytes ( - NULL, //the allocator - (const UInt8*)szUnConv, - strlen(szUnConv), - m_encoding, - false //no BOM/external representation - ); - - wxASSERT(theString); - - size_t nOutLength = CFStringGetLength(theString); - - if (szOut == NULL) - { - CFRelease(theString); - return nOutLength; - } - - CFRange theRange = { 0, nOutSize }; - -#if SIZEOF_WCHAR_T == 4 - UniChar* szUniCharBuffer = new UniChar[nOutSize]; -#endif - - CFStringGetCharacters(theString, theRange, szUniCharBuffer); - - CFRelease(theString); - - szUniCharBuffer[nOutLength] = '\0'; - -#if SIZEOF_WCHAR_T == 4 - wxMBConvUTF16 converter; - converter.MB2WC( szOut, (const char*)szUniCharBuffer, nOutSize ); - delete [] szUniCharBuffer; -#endif - - return nOutLength; - } - - size_t WC2MB(char *szOut, const wchar_t *szUnConv, size_t nOutSize) const - { - wxASSERT(szUnConv); - - size_t nRealOutSize; - size_t nBufSize = wxWcslen(szUnConv); - UniChar* szUniBuffer = (UniChar*) szUnConv; - -#if SIZEOF_WCHAR_T == 4 - 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); -#endif - - CFStringRef theString = CFStringCreateWithCharactersNoCopy( - NULL, //allocator - szUniBuffer, - nBufSize, - kCFAllocatorNull //deallocator - we want to deallocate it ourselves - ); - - wxASSERT(theString); - - //Note that CER puts a BOM when converting to unicode - //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); - - nRealOutSize = CFStringGetLength(theString) + 1; - } - else - { - 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 - ); - } - - CFRelease(theString); - -#if SIZEOF_WCHAR_T == 4 - delete[] szUniBuffer; -#endif - - return nRealOutSize - 1; - } - - virtual wxMBConv *Clone() const { return new wxMBConv_cocoa(*this); } - - bool IsOk() const - { - return m_encoding != kCFStringEncodingInvalidId && - CFStringIsEncodingAvailable(m_encoding); - } - -private: - CFStringEncoding m_encoding ; -}; - -#endif // defined(__WXCOCOA__) - -// ============================================================================ -// Mac conversion classes -// ============================================================================ - -#if defined(__WXMAC__) && defined(TARGET_CARBON) - -class wxMBConv_mac : public wxMBConv -{ -public: - wxMBConv_mac() - { - Init(CFStringGetSystemEncoding()) ; - } - - wxMBConv_mac(const wxMBConv_mac& conv) - { - Init(conv.m_char_encoding); - } - -#if wxUSE_FONTMAP - wxMBConv_mac(const char* name) - { - Init( wxMacGetSystemEncFromFontEnc( wxFontMapperBase::Get()->CharsetToEncoding(name, false) ) ); - } -#endif - - wxMBConv_mac(wxFontEncoding encoding) - { - Init( wxMacGetSystemEncFromFontEnc(encoding) ); - } - - virtual ~wxMBConv_mac() - { - OSStatus status = noErr ; - if (m_MB2WC_converter) - status = TECDisposeConverter(m_MB2WC_converter); - if (m_WC2MB_converter) - status = TECDisposeConverter(m_WC2MB_converter); - } - - void Init( TextEncodingBase encoding,TextEncodingVariant encodingVariant = kTextEncodingDefaultVariant , - TextEncodingFormat encodingFormat = kTextEncodingDefaultFormat) - { - m_MB2WC_converter = NULL ; - m_WC2MB_converter = NULL ; - m_char_encoding = CreateTextEncoding(encoding, encodingVariant, encodingFormat) ; - m_unicode_encoding = CreateTextEncoding(kTextEncodingUnicodeDefault, 0, kUnicode16BitFormat) ; - } - - virtual void CreateIfNeeded() const - { - if ( m_MB2WC_converter == NULL && m_WC2MB_converter == NULL ) - { - OSStatus status = noErr ; - status = TECCreateConverter(&m_MB2WC_converter, - m_char_encoding, - m_unicode_encoding); - wxASSERT_MSG( status == noErr , _("Unable to create TextEncodingConverter")) ; - status = TECCreateConverter(&m_WC2MB_converter, - m_unicode_encoding, - m_char_encoding); - wxASSERT_MSG( status == noErr , _("Unable to create TextEncodingConverter")) ; - } - } - - size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const - { - CreateIfNeeded() ; - OSStatus status = noErr ; - ByteCount byteOutLen ; - ByteCount byteInLen = strlen(psz) + 1; - wchar_t *tbuf = NULL ; - UniChar* ubuf = NULL ; - size_t res = 0 ; - - if (buf == NULL) - { - // Apple specs say at least 32 - n = wxMax( 32, byteInLen ) ; - tbuf = (wchar_t*) malloc( n * SIZEOF_WCHAR_T ) ; - } - - ByteCount byteBufferLen = n * sizeof( UniChar ) ; - -#if SIZEOF_WCHAR_T == 4 - ubuf = (UniChar*) malloc( byteBufferLen + 2 ) ; -#else - ubuf = (UniChar*) (buf ? buf : tbuf) ; -#endif - - status = TECConvertText( - m_MB2WC_converter, (ConstTextPtr) psz, byteInLen, &byteInLen, - (TextPtr) ubuf, byteBufferLen, &byteOutLen); - -#if SIZEOF_WCHAR_T == 4 - // 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 ; - wxMBConvUTF16 converter ; - res = converter.MB2WC( (buf ? buf : tbuf), (const char*)ubuf, n ) ; - free( ubuf ) ; -#else - res = byteOutLen / sizeof( UniChar ) ; -#endif - - if ( buf == NULL ) - free(tbuf) ; - - if ( buf && res < n) - buf[res] = 0; - - return res ; - } - - size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const - { - CreateIfNeeded() ; - OSStatus status = noErr ; - ByteCount byteOutLen ; - ByteCount byteInLen = wxWcslen(psz) * SIZEOF_WCHAR_T ; - - char *tbuf = NULL ; - - if (buf == NULL) - { - // 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 - wxMBConvUTF16 converter ; - size_t unicharlen = converter.WC2MB( NULL, psz, 0 ) ; - byteInLen = unicharlen ; - ubuf = (UniChar*) malloc( byteInLen + 2 ) ; - converter.WC2MB( (char*) ubuf, psz, unicharlen + 2 ) ; -#else - ubuf = (UniChar*) psz ; -#endif - - status = TECConvertText( - m_WC2MB_converter, (ConstTextPtr) ubuf, byteInLen, &byteInLen, - (TextPtr) (buf ? buf : tbuf), byteBufferLen, &byteOutLen); - -#if SIZEOF_WCHAR_T == 4 - free( ubuf ) ; -#endif - - if ( buf == NULL ) - free(tbuf) ; - - 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) == wxCONV_FAILED || - 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 wxCONV_FAILED; - } - } - - return res ; - } - - virtual wxMBConv *Clone() const { return new wxMBConv_mac(*this); } - - bool IsOk() const - { - CreateIfNeeded() ; - return m_MB2WC_converter != NULL && m_WC2MB_converter != NULL; - } - -protected : - mutable TECObjectRef m_MB2WC_converter; - mutable TECObjectRef m_WC2MB_converter; - - TextEncodingBase m_char_encoding; - TextEncodingBase m_unicode_encoding; -}; - -// MB is decomposed (D) normalized UTF8 - -class wxMBConv_macUTF8D : public wxMBConv_mac -{ -public : - wxMBConv_macUTF8D() - { - Init( kTextEncodingUnicodeDefault , kUnicodeNoSubset , kUnicodeUTF8Format ) ; - m_uni = NULL; - m_uniBack = NULL ; - } - - virtual ~wxMBConv_macUTF8D() - { - if (m_uni!=NULL) - DisposeUnicodeToTextInfo(&m_uni); - if (m_uniBack!=NULL) - DisposeUnicodeToTextInfo(&m_uniBack); - } - - size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const - { - CreateIfNeeded() ; - OSStatus status = noErr ; - ByteCount byteOutLen ; - ByteCount byteInLen = wxWcslen(psz) * SIZEOF_WCHAR_T ; - - char *tbuf = NULL ; - - if (buf == NULL) - { - // 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 - wxMBConvUTF16 converter ; - size_t unicharlen = converter.WC2MB( NULL, psz, 0 ) ; - byteInLen = unicharlen ; - ubuf = (UniChar*) malloc( byteInLen + 2 ) ; - converter.WC2MB( (char*) ubuf, psz, unicharlen + 2 ) ; -#else - ubuf = (UniChar*) psz ; -#endif - - // ubuf is a non-decomposed UniChar buffer - - ByteCount dcubuflen = byteInLen * 2 + 2 ; - ByteCount dcubufread , dcubufwritten ; - UniChar *dcubuf = (UniChar*) malloc( dcubuflen ) ; - - ConvertFromUnicodeToText( m_uni , byteInLen , ubuf , - kUnicodeDefaultDirectionMask, 0, NULL, NULL, NULL, dcubuflen , &dcubufread , &dcubufwritten , dcubuf ) ; - - // we now convert that decomposed buffer into UTF8 - - status = TECConvertText( - m_WC2MB_converter, (ConstTextPtr) dcubuf, dcubufwritten, &dcubufread, - (TextPtr) (buf ? buf : tbuf), byteBufferLen, &byteOutLen); - - free( dcubuf ); - -#if SIZEOF_WCHAR_T == 4 - free( ubuf ) ; -#endif - - if ( buf == NULL ) - free(tbuf) ; - - size_t res = byteOutLen ; - if ( buf && res < n) - { - buf[res] = 0; - // don't test for round-trip fidelity yet, we cannot guarantee it yet - } - - return res ; - } - - size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const - { - CreateIfNeeded() ; - OSStatus status = noErr ; - ByteCount byteOutLen ; - ByteCount byteInLen = strlen(psz) + 1; - wchar_t *tbuf = NULL ; - UniChar* ubuf = NULL ; - size_t res = 0 ; - - if (buf == NULL) - { - // Apple specs say at least 32 - n = wxMax( 32, byteInLen ) ; - tbuf = (wchar_t*) malloc( n * SIZEOF_WCHAR_T ) ; - } - - ByteCount byteBufferLen = n * sizeof( UniChar ) ; - -#if SIZEOF_WCHAR_T == 4 - ubuf = (UniChar*) malloc( byteBufferLen + 2 ) ; -#else - ubuf = (UniChar*) (buf ? buf : tbuf) ; -#endif - - ByteCount dcubuflen = byteBufferLen * 2 + 2 ; - ByteCount dcubufread , dcubufwritten ; - UniChar *dcubuf = (UniChar*) malloc( dcubuflen ) ; - - status = TECConvertText( - m_MB2WC_converter, (ConstTextPtr) psz, byteInLen, &byteInLen, - (TextPtr) dcubuf, dcubuflen, &byteOutLen); - // 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 - dcubuf[byteOutLen / sizeof( UniChar ) ] = 0 ; - - // now from the decomposed UniChar to properly composed uniChar - ConvertFromUnicodeToText( m_uniBack , byteOutLen , dcubuf , - kUnicodeDefaultDirectionMask, 0, NULL, NULL, NULL, dcubuflen , &dcubufread , &dcubufwritten , ubuf ) ; - - free( dcubuf ); - byteOutLen = dcubufwritten ; - ubuf[byteOutLen / sizeof( UniChar ) ] = 0 ; - - -#if SIZEOF_WCHAR_T == 4 - wxMBConvUTF16 converter ; - res = converter.MB2WC( (buf ? buf : tbuf), (const char*)ubuf, n ) ; - free( ubuf ) ; -#else - res = byteOutLen / sizeof( UniChar ) ; -#endif - - if ( buf == NULL ) - free(tbuf) ; - - if ( buf && res < n) - buf[res] = 0; - - return res ; - } - - virtual void CreateIfNeeded() const - { - wxMBConv_mac::CreateIfNeeded() ; - if ( m_uni == NULL ) - { - m_map.unicodeEncoding = CreateTextEncoding(kTextEncodingUnicodeDefault, - kUnicodeNoSubset, kTextEncodingDefaultFormat); - m_map.otherEncoding = CreateTextEncoding(kTextEncodingUnicodeDefault, - kUnicodeCanonicalDecompVariant, kTextEncodingDefaultFormat); - m_map.mappingVersion = kUnicodeUseLatestMapping; - - OSStatus err = CreateUnicodeToTextInfo(&m_map, &m_uni); - wxASSERT_MSG( err == noErr , _(" Couldn't create the UnicodeConverter")) ; - - m_map.unicodeEncoding = CreateTextEncoding(kTextEncodingUnicodeDefault, - kUnicodeNoSubset, kTextEncodingDefaultFormat); - m_map.otherEncoding = CreateTextEncoding(kTextEncodingUnicodeDefault, - kUnicodeCanonicalCompVariant, kTextEncodingDefaultFormat); - m_map.mappingVersion = kUnicodeUseLatestMapping; - err = CreateUnicodeToTextInfo(&m_map, &m_uniBack); - wxASSERT_MSG( err == noErr , _(" Couldn't create the UnicodeConverter")) ; - } - } -protected : - mutable UnicodeToTextInfo m_uni; - mutable UnicodeToTextInfo m_uniBack; - mutable UnicodeMapping m_map; -}; -#endif // defined(__WXMAC__) && defined(TARGET_CARBON) // ============================================================================ // wxEncodingConverter based conversion classes @@ -3096,7 +2576,10 @@ class wxMBConv_wxwin : public wxMBConv private: void Init() { - m_ok = m2w.Init(m_enc, wxFONTENCODING_UNICODE) && + // Refuse to use broken wxEncodingConverter code for Mac-specific encodings. + // The wxMBConv_cf class does a better job. + m_ok = (m_enc < wxFONTENCODING_MACMIN || m_enc > wxFONTENCODING_MACMAX) && + m2w.Init(m_enc, wxFONTENCODING_UNICODE) && w2m.Init(wxFONTENCODING_UNICODE, m_enc); } @@ -3268,7 +2751,7 @@ void wxCSConv::SetName(const char *charset) { if (charset) { - m_name = strdup(charset); + m_name = wxStrdup(charset); m_deferred = true; } } @@ -3391,35 +2874,17 @@ wxMBConv *wxCSConv::DoCreate() const } #endif // wxHAVE_WIN32_MB2WC -#if defined(__WXMAC__) +#ifdef __DARWIN__ { // 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; - - delete conv; - } - } -#endif - -#if defined(__WXCOCOA__) - { - 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); + wxMBConv_cf *conv = m_name ? new wxMBConv_cf(m_name) + : new wxMBConv_cf(m_encoding); #else - wxMBConv_cocoa *conv = new wxMBConv_cocoa(m_encoding); + wxMBConv_cf *conv = new wxMBConv_cf(m_encoding); #endif if ( conv->IsOk() ) @@ -3428,7 +2893,8 @@ wxMBConv *wxCSConv::DoCreate() const delete conv; } } -#endif +#endif // __DARWIN__ + // step (2) wxFontEncoding enc = m_encoding; #if wxUSE_FONTMAP @@ -3708,13 +3174,11 @@ wxCharBuffer wxSafeConvertWX2MB(const wchar_t *ws) #ifdef __WINDOWS__ WX_DEFINE_GLOBAL_CONV2(wxMBConv, wxMBConv_win32, wxConvLibc, wxEMPTY_PARAMETER_VALUE); -#elif defined(__WXMAC__) && !defined(__MACH__) - WX_DEFINE_GLOBAL_CONV2(wxMBConv, wxMBConv_mac, wxConvLibc, wxEMPTY_PARAMETER_VALUE); #else WX_DEFINE_GLOBAL_CONV2(wxMBConv, wxMBConvLibc, wxConvLibc, wxEMPTY_PARAMETER_VALUE); #endif -WX_DEFINE_GLOBAL_CONV(wxMBConvUTF8, wxConvUTF8, wxEMPTY_PARAMETER_VALUE); +WX_DEFINE_GLOBAL_CONV(wxMBConvStrictUTF8, wxConvUTF8, wxEMPTY_PARAMETER_VALUE); WX_DEFINE_GLOBAL_CONV(wxMBConvUTF7, wxConvUTF7, wxEMPTY_PARAMETER_VALUE); WX_DEFINE_GLOBAL_CONV(wxCSConv, wxConvLocal, (wxFONTENCODING_SYSTEM)); @@ -3723,19 +3187,18 @@ WX_DEFINE_GLOBAL_CONV(wxCSConv, wxConvISO8859_1, (wxFONTENCODING_ISO8859_1)); WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent = wxGet_wxConvLibcPtr(); WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvUI = wxGet_wxConvLocalPtr(); -#if defined(__WXMAC__) && defined(TARGET_CARBON) -static wxMBConv_macUTF8D wxConvMacUTF8DObj; +#ifdef __DARWIN__ +// The xnu kernel always communicates file paths in decomposed UTF-8. +// WARNING: Are we sure that CFString's conversion will cause decomposition? +static wxMBConv_cf wxConvMacUTF8DObj(wxFONTENCODING_UTF8); #endif + WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName = -#ifdef __WXOSX__ -#if defined(__WXMAC__) && defined(TARGET_CARBON) +#ifdef __DARWIN__ &wxConvMacUTF8DObj; -#else - wxGet_wxConvUTF8Ptr(); -#endif -#else // !__WXOSX__ +#else // !__DARWIN__ wxGet_wxConvLibcPtr(); -#endif // __WXOSX__/!__WXOSX__ +#endif // __DARWIN__/!__DARWIN__ #else // !wxUSE_WCHAR_T