X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/6ff49cbcd4d66b958014176594b9be1796bd243a..69659fd770f615210efac4b4fa741b3ad6223616:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index fb6b7b6f9c..59ea721b71 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -57,23 +57,9 @@ #include "wx/fontmap.h" #ifdef __DARWIN__ -#include -#include - -#include "wx/mac/corefoundation/cfref.h" +#include "wx/mac/corefoundation/private/strconv_cf.h" #endif //def __DARWIN__ -#ifdef __WXMAC__ -#ifndef __DARWIN__ -#include -#include -#include -#endif - -// includes Mac headers -#include "wx/mac/private.h" -#endif - #define TRACE_STRCONV _T("strconv") @@ -2303,868 +2289,6 @@ private: #endif // wxHAVE_WIN32_MB2WC -// ============================================================================ -// CoreFoundation conversion classes -// ============================================================================ - -#ifdef __DARWIN__ - -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 ; -/* Don't support conversion to/from UTF16 as wxWidgets can do this better. - * In particular, ToWChar would fail miserably using strlen on an input UTF16. - 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 ; -} - -#if MAC_OS_X_VERSION_MAX_ALLOWED >= MAC_OS_X_VERSION_10_4 -// Provide a constant for the wchat_t encoding used by the host platform. -#ifdef WORDS_BIGENDIAN - static const CFStringEncoding wxCFStringEncodingWcharT = kCFStringEncodingUTF32BE; -#else - static const CFStringEncoding wxCFStringEncodingWcharT = kCFStringEncodingUTF32LE; -#endif - -#endif /* MAC_OS_X_VERSION_MAX_ALLOWED >= MAC_OS_X_VERSION_10_4 */ - -class wxMBConv_cf : public wxMBConv -{ -public: - wxMBConv_cf() - { - Init(CFStringGetSystemEncoding()) ; - } - - wxMBConv_cf(const wxMBConv_cf& conv) - { - m_encoding = conv.m_encoding; - } - -#if wxUSE_FONTMAP - wxMBConv_cf(const char* name) - { - Init( wxCFStringEncFromFontEnc(wxFontMapperBase::Get()->CharsetToEncoding(name, false) ) ) ; - } -#endif - - wxMBConv_cf(wxFontEncoding encoding) - { - Init( wxCFStringEncFromFontEnc(encoding) ); - } - - virtual ~wxMBConv_cf() - { - } - - void Init( CFStringEncoding encoding) - { - m_encoding = encoding ; - } - - virtual size_t ToWChar(wchar_t * dst, size_t dstSize, const char * src, size_t srcSize = wxNO_LEN) const - { - wxCHECK(src, wxCONV_FAILED); - - /* NOTE: This is wrong if the source encoding has an element size - * other than char (e.g. it's kCFStringEncodingUnicode) - * If the user specifies it, it's presumably right though. - * Right now we don't support UTF-16 in anyway since wx can do a better job. - */ - if(srcSize == wxNO_LEN) - srcSize = strlen(src) + 1; - - // First create the temporary CFString - wxCFRef theString( CFStringCreateWithBytes ( - NULL, //the allocator - (const UInt8*)src, - srcSize, - m_encoding, - false //no BOM/external representation - )); - - wxCHECK(theString != NULL, wxCONV_FAILED); - - /* NOTE: The string content includes the NULL element if the source string did - * That means we have to do nothing special because the destination will have - * the NULL element iff the source did and the NULL element will be included - * in the count iff it was included in the source count. - */ - - -/* If we're compiling against Tiger headers we can support direct conversion - * to UTF32. If we are then run against a pre-Tiger system, the encoding - * won't be available so we'll defer to the string->UTF-16->UTF-32 conversion. - */ -#if MAC_OS_X_VERSION_MAX_ALLOWED >= MAC_OS_X_VERSION_10_4 - if(CFStringIsEncodingAvailable(wxCFStringEncodingWcharT)) - { - CFRange fullStringRange = CFRangeMake(0, CFStringGetLength(theString)); - CFIndex usedBufLen; - - CFIndex charsConverted = CFStringGetBytes( - theString, - fullStringRange, - wxCFStringEncodingWcharT, - 0, - false, - // if dstSize is 0 then pass NULL to get required length in usedBufLen - dstSize != 0?(UInt8*)dst:NULL, - dstSize * sizeof(wchar_t), - &usedBufLen); - - // charsConverted is > 0 iff conversion succeeded - if(charsConverted <= 0) - return wxCONV_FAILED; - - /* usedBufLen is the number of bytes written, so we divide by - * sizeof(wchar_t) to get the number of elements written. - */ - wxASSERT( (usedBufLen % sizeof(wchar_t)) == 0 ); - - // CFStringGetBytes does exactly the right thing when buffer - // pointer is NULL and returns the number of bytes required - return usedBufLen / sizeof(wchar_t); - } - else -#endif /* MAC_OS_X_VERSION_MAX_ALLOWED >= MAC_OS_X_VERSION_10_4 */ - { - // NOTE: Includes NULL iff source did - /* NOTE: This is an approximation. The eventual UTF-32 will - * possibly have less elements but certainly not more. - */ - size_t returnSize = CFStringGetLength(theString); - - if (dstSize == 0 || dst == NULL) - { - return returnSize; - } - - // Convert the entire string.. too hard to figure out how many UTF-16 we'd need - // for an undersized UTF-32 destination buffer. - CFRange fullStringRange = CFRangeMake(0, CFStringGetLength(theString)); - UniChar *szUniCharBuffer = new UniChar[fullStringRange.length]; - - CFStringGetCharacters(theString, fullStringRange, szUniCharBuffer); - - wxMBConvUTF16 converter; - returnSize = converter.ToWChar( dst, dstSize, (const char*)szUniCharBuffer, fullStringRange.length ); - delete [] szUniCharBuffer; - - return returnSize; - } - // NOTREACHED - } - - virtual size_t FromWChar(char *dst, size_t dstSize, const wchar_t *src, size_t srcSize) const - { - wxCHECK(src, wxCONV_FAILED); - - if(srcSize == wxNO_LEN) - srcSize = wxStrlen(src) + 1; - - // Temporary CFString - wxCFRef theString; - -/* If we're compiling against Tiger headers we can support direct conversion - * from UTF32. If we are then run against a pre-Tiger system, the encoding - * won't be available so we'll defer to the UTF-32->UTF-16->string conversion. - */ -#if MAC_OS_X_VERSION_MAX_ALLOWED >= MAC_OS_X_VERSION_10_4 - if(CFStringIsEncodingAvailable(wxCFStringEncodingWcharT)) - { - theString = wxCFRef(CFStringCreateWithBytes( - kCFAllocatorDefault, - (UInt8*)src, - srcSize * sizeof(wchar_t), - wxCFStringEncodingWcharT, - false)); - } - else -#endif /* MAC_OS_X_VERSION_MAX_ALLOWED >= MAC_OS_X_VERSION_10_4 */ - { - wxMBConvUTF16 converter; - size_t cbUniBuffer = converter.FromWChar( NULL, 0, src, srcSize ); - wxASSERT(cbUniBuffer % sizeof(UniChar)); - - // Will be free'd by kCFAllocatorMalloc when CFString is released - UniChar *tmpUniBuffer = (UniChar*)malloc(cbUniBuffer); - - cbUniBuffer = converter.FromWChar( (char*) tmpUniBuffer, cbUniBuffer, src, srcSize ); - wxASSERT(cbUniBuffer % sizeof(UniChar)); - - theString = wxCFRef(CFStringCreateWithCharactersNoCopy( - kCFAllocatorDefault, - tmpUniBuffer, - cbUniBuffer / sizeof(UniChar), - kCFAllocatorMalloc - )); - - } - - wxCHECK(theString != NULL, wxCONV_FAILED); - - CFIndex usedBufLen; - - CFIndex charsConverted = CFStringGetBytes( - theString, - CFRangeMake(0, CFStringGetLength(theString)), - m_encoding, - 0, // FAIL on unconvertible characters - false, // not an external representation - // if dstSize is 0 then pass NULL to get required length in usedBufLen - (dstSize != 0)?(UInt8*)dst:NULL, - dstSize, - &usedBufLen - ); - - // charsConverted is > 0 iff conversion succeeded - if(charsConverted <= 0) - return wxCONV_FAILED; - - return usedBufLen; - } - - virtual wxMBConv *Clone() const { return new wxMBConv_cf(*this); } - - bool IsOk() const - { - return m_encoding != kCFStringEncodingInvalidId && - CFStringIsEncodingAvailable(m_encoding); - } - -private: - CFStringEncoding m_encoding ; -}; - -#endif // __DARWIN__ - -// ============================================================================ -// Mac conversion classes -// ============================================================================ - -/* Although we are in the base library we currently have this wxMac - * conditional. This is not generally good but fortunately does not affect - * the ABI of the base library, only what encodings might work. - * It does mean that a wxBase built as part of wxMac has slightly more support - * than one built for wxCocoa or even wxGtk. - */ -#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 @@ -3177,7 +2301,10 @@ 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); } @@ -3472,26 +2599,6 @@ wxMBConv *wxCSConv::DoCreate() const } #endif // wxHAVE_WIN32_MB2WC -#if defined(__WXMAC__) - { - // 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 - #ifdef __DARWIN__ { // leave UTF16 and UTF32 to the built-ins of wx @@ -3792,8 +2899,6 @@ 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 @@ -3807,19 +2912,18 @@ 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