X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/5c3c8676d6a99dc3e91ce1de6971dc44279aac38..4b524c27cceda99907dec63f9224270ed5dd617d:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 792d2caa50..a43f43550f 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -55,6 +55,9 @@ #include #include #include +#ifdef HAVE_LANGINFO_H + #include +#endif #if defined(__WIN32__) && !defined(__WXMICROWIN__) #define wxHAVE_WIN32_MB2WC @@ -278,7 +281,7 @@ const wxWCharBuffer wxMBConv::cMB2WC(const char *szString, size_t nStringLen, si //success - return actual length and the buffer *pOutSize = nActualLength; - return theBuffer; + return theBuffer; } const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *szString, size_t nStringLen, size_t* pOutSize) const @@ -313,7 +316,7 @@ const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *szString, size_t nStringLen, //Increase the actual length (+1 for current null character) nActualLength += nLen + 1; - + //if buffer too big, realloc the buffer if (nActualLength > (nCurrentSize+1)) { @@ -340,7 +343,7 @@ const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *szString, size_t nStringLen, //success - return actual length and the buffer *pOutSize = nActualLength; - return theBuffer; + return theBuffer; } // ---------------------------------------------------------------------------- @@ -356,8 +359,63 @@ size_t wxMBConvLibc::WC2MB(char *buf, const wchar_t *psz, size_t n) const { return wxWC2MB(buf, psz, n); } + +#ifdef __UNIX__ + +// ---------------------------------------------------------------------------- +// wxConvBrokenFileNames +// ---------------------------------------------------------------------------- + +wxConvBrokenFileNames::wxConvBrokenFileNames() +{ + // decide which conversion to use for the file names + + // (1) this variable exists for the sole purpose of specifying the encoding + // of the filenames for GTK+ programs, so use it if it is set + const wxChar *encName = wxGetenv(_T("G_FILENAME_ENCODING")); + if ( encName ) + { + m_conv = new wxCSConv(encName); + } + else // no G_FILENAME_ENCODING + { + // (2) if a non default locale is set, assume that the user wants his + // filenames in this locale too + switch ( wxLocale::GetSystemEncoding() ) + { + default: + m_conv = new wxMBConvLibc; + break; + + // (3) finally use UTF-8 by default + case wxFONTENCODING_SYSTEM: + case wxFONTENCODING_UTF8: + m_conv = new wxMBConvUTF8(wxMBConvUTF8::MAP_INVALID_UTF8_TO_OCTAL); + break; + } + } +} + +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 @@ -510,7 +568,7 @@ size_t wxMBConvUTF7::WC2MB(char *buf, const wchar_t *psz, size_t n) const } #ifndef WC_UTF16 else if (((wxUint32)cc) > 0xffff) - { + { // no surrogate pair generation (yet?) return (size_t)-1; } @@ -567,12 +625,19 @@ size_t wxMBConvUTF7::WC2MB(char *buf, const wchar_t *psz, size_t n) const 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++) @@ -590,7 +655,7 @@ 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 { @@ -598,32 +663,93 @@ 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; + } } } } @@ -632,6 +758,11 @@ 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; @@ -646,36 +777,59 @@ 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 ) { - // plain ASCII char if (buf) - *buf++ = (char) cc; + *buf++ = (char)(cc - wxUnicodePUA); 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; @@ -1371,7 +1525,7 @@ size_t wxMBConv_iconv::WC2MB(char *buf, const wchar_t *psz, size_t n) const // 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; @@ -1899,7 +2053,7 @@ public: #if wxUSE_FONTMAP wxMBConv_cocoa(const wxChar* name) { - Init( wxCFStringEncFromFontEnc(wxFontMapper::Get()->CharsetToEncoding(name, false) ) ) ; + Init( wxCFStringEncFromFontEnc(wxFontMapperBase::Get()->CharsetToEncoding(name, false) ) ) ; } #endif @@ -1944,9 +2098,9 @@ public: #if SIZEOF_WCHAR_T == 4 UniChar* szUniCharBuffer = new UniChar[nOutSize]; #endif - + CFStringGetCharacters(theString, theRange, szUniCharBuffer); - + CFRelease(theString); szUniCharBuffer[nOutLength] = '\0' ; @@ -1956,14 +2110,14 @@ public: 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; @@ -1991,7 +2145,7 @@ public: { if (szOut != NULL) CFStringGetCharacters(theString, CFRangeMake(0, nOutSize - 1), (UniChar*) szOut); - + nRealOutSize = CFStringGetLength(theString) + 1; } else @@ -2004,7 +2158,7 @@ public: //0 tells CFString to return NULL if it meets such a character false, //not an external representation (UInt8*) szOut, - nOutSize, + nOutSize, (CFIndex*) &nRealOutSize ); } @@ -2020,7 +2174,7 @@ public: bool IsOk() const { - return m_encoding != kCFStringEncodingInvalidId && + return m_encoding != kCFStringEncodingInvalidId && CFStringIsEncodingAvailable(m_encoding); } @@ -2047,7 +2201,7 @@ public: #if wxUSE_FONTMAP wxMBConv_mac(const wxChar* name) { - Init( wxMacGetSystemEncFromFontEnc(wxFontMapper::Get()->CharsetToEncoding(name, false) ) ) ; + Init( wxMacGetSystemEncFromFontEnc(wxFontMapperBase::Get()->CharsetToEncoding(name, false) ) ) ; } #endif @@ -2158,7 +2312,7 @@ public: 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); @@ -2210,7 +2364,7 @@ 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; @@ -2369,7 +2523,7 @@ 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); @@ -2397,7 +2551,7 @@ wxMBConv *wxCSConv::DoCreate() const #if defined(__WXMAC__) { // leave UTF16 and UTF32 to the built-ins of wx - if ( m_name || ( m_encoding < wxFONTENCODING_UTF16BE || + if ( m_name || ( m_encoding < wxFONTENCODING_UTF16BE || ( m_encoding >= wxFONTENCODING_MACMIN && m_encoding <= wxFONTENCODING_MACMAX ) ) ) { @@ -2440,7 +2594,7 @@ 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 @@ -2496,7 +2650,7 @@ 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 @@ -2601,9 +2755,9 @@ WXDLLIMPEXP_DATA_BASE(wxMBConvUTF8&) wxConvUTF8 = wxConvUTF8Obj; WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent = &wxConvLibcObj; WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName = & #ifdef __WXOSX__ - wxConvUTF8Obj; + wxConvUTF8Obj; #else - wxConvLibcObj; + wxConvLibcObj; #endif