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;
//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
//Increase the actual length (+1 for current null character)
nActualLength += nLen + 1;
-
+
//if buffer too big, realloc the buffer
if (nActualLength > (nCurrentSize+1))
{
//success - return actual length and the buffer
*pOutSize = nActualLength;
- return theBuffer;
+ return theBuffer;
}
// ----------------------------------------------------------------------------
return wxWC2MB(buf, psz, n);
}
+#ifdef __UNIX__
+
// ----------------------------------------------------------------------------
-// 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 in disk.
+// wxConvBrokenFileNames
// ----------------------------------------------------------------------------
-class wxConvBrokenFileNames: public wxMBConvLibc
+wxConvBrokenFileNames::wxConvBrokenFileNames()
{
-public:
- 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;
-};
+ // decide which conversion to use for the file names
-size_t wxConvBrokenFileNames::MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const
-{
-#if 0
- if (we find some invalid characters)
+ // (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
+ wxString encName(wxGetenv(_T("G_FILENAME_ENCODING")));
+ encName.MakeUpper();
+ if ( !encName.empty() && encName != _T("UTF-8") && encName != _T("UTF8") )
{
- Convert to Unicode range.
+ m_conv = new wxCSConv(encName);
+ }
+ else // no G_FILENAME_ENCODING
+ {
+ if ( encName.empty() )
+ encName = wxLocale::GetSystemEncodingName().Upper();
+
+ // (2) if a non default locale is set, assume that the user wants his
+ // filenames in this locale too
+ if ( !encName.empty() && encName != _T("UTF-8") && encName != _T("UTF8") )
+ {
+ wxSetEnv(_T("G_FILENAME_ENCODING"), encName);
+ m_conv = new wxMBConvLibc;
+ }
+ else
+ {
+ // (3) finally use UTF-8 by default
+ m_conv = new wxMBConvUTF8(wxMBConvUTF8::MAP_INVALID_UTF8_TO_OCTAL);
+ }
}
- else
-#endif
- return wxMBConvLibc::MB2WC( outputBuf, psz, outputSize );
}
-size_t wxConvBrokenFileNames::WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const
+size_t
+wxConvBrokenFileNames::MB2WC(wchar_t *outputBuf,
+ const char *psz,
+ size_t outputSize) const
{
-#if 0
- Convert back from Unicode range.
-#endif
- return wxMBConvLibc::WC2MB( outputBuf, psz, outputSize );
+ 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
}
#ifndef WC_UTF16
else if (((wxUint32)cc) > 0xffff)
- {
+ {
// no surrogate pair generation (yet?)
return (size_t)-1;
}
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++)
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
{
if (!cnt)
{
// invalid UTF-8 sequence
- return (size_t)-1;
+ invalid = true;
}
else
{
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;
+ }
}
}
}
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;
#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 )
+ {
+ if (buf)
+ *buf++ = (char)(cc - wxUnicodePUA);
+ len++;
+ }
+ else if ( (m_options & MAP_INVALID_UTF8_TO_OCTAL)
+ && cc == L'\\' && psz[0] == L'\\' )
{
- // plain ASCII char
if (buf)
- *buf++ = (char) cc;
+ *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 && (len<n)) *buf = 0;
+ if (buf && (len<n))
+ *buf = 0;
return len;
}
-
-
-
// ----------------------------------------------------------------------------
// UTF-16
// ----------------------------------------------------------------------------
// as MB<->WC 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;
// 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;
#if SIZEOF_WCHAR_T == 4
UniChar* szUniCharBuffer = new UniChar[nOutSize];
#endif
-
+
CFStringGetCharacters(theString, theRange, szUniCharBuffer);
-
+
CFRelease(theString);
szUniCharBuffer[nOutLength] = '\0' ;
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 (szOut != NULL)
CFStringGetCharacters(theString, CFRangeMake(0, nOutSize - 1), (UniChar*) szOut);
-
+
nRealOutSize = CFStringGetLength(theString) + 1;
}
else
//0 tells CFString to return NULL if it meets such a character
false, //not an external representation
(UInt8*) szOut,
- nOutSize,
+ nOutSize,
(CFIndex*) &nRealOutSize
);
}
bool IsOk() const
{
- return m_encoding != kCFStringEncodingInvalidId &&
+ return m_encoding != kCFStringEncodingInvalidId &&
CFStringIsEncodingAvailable(m_encoding);
}
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);
#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 ) ) )
{
WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent = &wxConvLibcObj;
WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName = &
#ifdef __WXOSX__
- wxConvUTF8Obj;
+ wxConvUTF8Obj;
#else
- wxConvLibcObj;
+ wxConvLibcObj;
#endif
projects / wxWidgets.git / blobdiff