X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/b1d547ebc5994e4e3837977812d8bb45cbdfa447..b9efe021b554fa3967d1442cf758435c5cd5ae8f:/src/common/strconv.cpp?ds=sidebyside diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 70493dbc12..aea1e1a18a 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -152,7 +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; @@ -270,15 +270,15 @@ const wxWCharBuffer wxMBConv::cMB2WC(const char *szString, size_t nStringLen, si } //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; } //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 +313,7 @@ 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)) { @@ -332,15 +332,15 @@ const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *szString, size_t nStringLen, } //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; } //success - return actual length and the buffer *pOutSize = nActualLength; - return theBuffer; + return theBuffer; } // ---------------------------------------------------------------------------- @@ -356,8 +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 @@ -403,7 +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))) @@ -493,8 +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 { @@ -512,7 +544,7 @@ size_t wxMBConvUTF7::WC2MB(char *buf, const wchar_t } #ifndef WC_UTF16 else if (((wxUint32)cc) > 0xffff) - { + { // no surrogate pair generation (yet?) return (size_t)-1; } @@ -569,12 +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++) @@ -585,6 +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 { @@ -592,7 +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 { @@ -600,32 +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; + } } } } @@ -634,6 +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; @@ -648,36 +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 ) { - // plain ASCII char if (buf) - *buf++ = (char) cc; + *buf++ = (char)(cc - wxUnicodePUA); len++; } + else if ( (m_options & MAP_INVALID_UTF8_TO_OCTAL) + && cc == L'\\' && psz[0] == L'\\' ) + { + if (buf) + *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; @@ -1373,7 +1518,7 @@ 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; @@ -1477,10 +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 @@ -1888,10 +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) { @@ -1934,9 +2091,9 @@ public: #if SIZEOF_WCHAR_T == 4 UniChar* szUniCharBuffer = new UniChar[nOutSize]; #endif - + CFStringGetCharacters(theString, theRange, szUniCharBuffer); - + CFRelease(theString); szUniCharBuffer[nOutLength] = '\0' ; @@ -1946,20 +2103,20 @@ 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; #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)) ; @@ -1981,7 +2138,7 @@ public: { if (szOut != NULL) CFStringGetCharacters(theString, CFRangeMake(0, nOutSize - 1), (UniChar*) szOut); - + nRealOutSize = CFStringGetLength(theString) + 1; } else @@ -1994,7 +2151,7 @@ public: //0 tells CFString to return NULL if it meets such a character false, //not an external representation (UInt8*) szOut, - nOutSize, + nOutSize, (CFIndex*) &nRealOutSize ); } @@ -2010,7 +2167,7 @@ public: bool IsOk() const { - return m_encoding != kCFStringEncodingInvalidId && + return m_encoding != kCFStringEncodingInvalidId && CFStringIsEncodingAvailable(m_encoding); } @@ -2034,10 +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) { @@ -2093,7 +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 @@ -2126,7 +2285,7 @@ public: 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 ) ; @@ -2146,7 +2305,7 @@ 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); @@ -2198,7 +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; @@ -2216,7 +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; } @@ -2224,7 +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; } @@ -2351,7 +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); @@ -2378,11 +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; @@ -2395,8 +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; @@ -2412,7 +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 @@ -2468,7 +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 @@ -2565,13 +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