X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/667e5b3e05b4769409203611fc7022f35780309c..ede3a6d68af66772b4f5f94208b4126bab566cc8:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 5ceede4225..db0147d6bc 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 +359,53 @@ size_t wxMBConvLibc::WC2MB(char *buf, const wchar_t *psz, size_t n) const { return wxWC2MB(buf, psz, n); } + // ---------------------------------------------------------------------------- -// UTF-7 +// wxConvBrokenFileNames is made for GTK2 in Unicode mode when +// files are accidentally written in an encoding which is not +// the system encoding. Typically, the system encoding will be +// UTF8 but there might be files stored in ISO8859-1 on disk. +// ---------------------------------------------------------------------------- + +class wxConvBrokenFileNames: public wxMBConvLibc +{ +public: + wxConvBrokenFileNames() : m_utf8conv(wxMBConvUTF8::MAP_INVALID_UTF8_TO_OCTAL) { } + virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const; + virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const; + inline bool UseUTF8() const; +private: + wxMBConvUTF8 m_utf8conv; +}; + +bool wxConvBrokenFileNames::UseUTF8() const +{ +#if defined HAVE_LANGINFO_H && defined CODESET + char *codeset = nl_langinfo(CODESET); + return strcmp(codeset, "UTF-8") == 0; +#else + return false; +#endif +} + +size_t wxConvBrokenFileNames::MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const +{ + if (UseUTF8()) + return m_utf8conv.MB2WC( outputBuf, psz, outputSize ); + else + return wxMBConvLibc::MB2WC( outputBuf, psz, outputSize ); +} + +size_t wxConvBrokenFileNames::WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const +{ + if (UseUTF8()) + return m_utf8conv.WC2MB( outputBuf, psz, outputSize ); + else + return wxMBConvLibc::WC2MB( outputBuf, psz, outputSize ); +} + +// ---------------------------------------------------------------------------- +// UTF-7 // ---------------------------------------------------------------------------- // Implementation (C) 2004 Fredrik Roubert @@ -567,12 +615,19 @@ 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 +645,7 @@ 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 +653,93 @@ 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++ = L'0' + n / 0100; + *buf++ = L'0' + (n % 0100) / 010; + *buf++ = L'0' + n % 010; + } + opsz++; + len += 4; + } + } + else // MAP_INVALID_UTF8_NOT + { + return (size_t)-1; + } } } } @@ -632,6 +748,11 @@ 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 +767,59 @@ 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 +1515,7 @@ 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; @@ -1896,10 +2040,12 @@ 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) { @@ -1942,9 +2088,9 @@ public: #if SIZEOF_WCHAR_T == 4 UniChar* szUniCharBuffer = new UniChar[nOutSize]; #endif - + CFStringGetCharacters(theString, theRange, szUniCharBuffer); - + CFRelease(theString); szUniCharBuffer[nOutLength] = '\0' ; @@ -1954,14 +2100,14 @@ 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; @@ -1989,7 +2135,7 @@ public: { if (szOut != NULL) CFStringGetCharacters(theString, CFRangeMake(0, nOutSize - 1), (UniChar*) szOut); - + nRealOutSize = CFStringGetLength(theString) + 1; } else @@ -2002,7 +2148,7 @@ public: //0 tells CFString to return NULL if it meets such a character false, //not an external representation (UInt8*) szOut, - nOutSize, + nOutSize, (CFIndex*) &nRealOutSize ); } @@ -2018,7 +2164,7 @@ public: bool IsOk() const { - return m_encoding != kCFStringEncodingInvalidId && + return m_encoding != kCFStringEncodingInvalidId && CFStringIsEncodingAvailable(m_encoding); } @@ -2042,10 +2188,12 @@ 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) { @@ -2154,7 +2302,7 @@ 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); @@ -2206,7 +2354,7 @@ 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; @@ -2365,7 +2513,7 @@ 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); @@ -2392,11 +2540,17 @@ 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; @@ -2409,8 +2563,12 @@ 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; @@ -2426,7 +2584,7 @@ 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 @@ -2482,7 +2640,7 @@ 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 @@ -2578,7 +2736,7 @@ static wxCSConv wxConvLocalObj(wxFONTENCODING_SYSTEM); static wxCSConv wxConvISO8859_1Obj(wxFONTENCODING_ISO8859_1); static wxMBConvUTF7 wxConvUTF7Obj; static wxMBConvUTF8 wxConvUTF8Obj; - +static wxConvBrokenFileNames wxConvBrokenFileNamesObj; WXDLLIMPEXP_DATA_BASE(wxMBConv&) wxConvLibc = wxConvLibcObj; WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvLocal = wxConvLocalObj; @@ -2586,6 +2744,15 @@ 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; +#elif __WXGTK20__ + wxConvBrokenFileNamesObj; +#else + wxConvLibcObj; +#endif + #else // !wxUSE_WCHAR_T