+size_t wxMBConvUTF7::ToWChar(wchar_t *dst, size_t dstLen,
+ const char *src, size_t srcLen) const
+{
+ DecoderState stateOrig,
+ *statePtr;
+ if ( srcLen == wxNO_LEN )
+ {
+ // convert the entire string, up to and including the trailing NUL
+ srcLen = strlen(src) + 1;
+
+ // when working on the entire strings we don't update nor use the shift
+ // state from the previous call
+ statePtr = &stateOrig;
+ }
+ else // when working with partial strings we do use the shift state
+ {
+ statePtr = wx_const_cast(DecoderState *, &m_stateDecoder);
+
+ // also save the old state to be able to rollback to it on error
+ stateOrig = m_stateDecoder;
+ }
+
+ // but to simplify the code below we use this variable in both cases
+ DecoderState& state = *statePtr;
+
+
+ // number of characters [which would have been] written to dst [if it were
+ // not NULL]
+ size_t len = 0;
+
+ const char * const srcEnd = src + srcLen;
+
+ while ( (src < srcEnd) && (!dst || (len < dstLen)) )
+ {
+ const unsigned char cc = *src++;
+
+ if ( state.IsShifted() )
+ {
+ const unsigned char dc = utf7unb64[cc];
+ if ( dc == 0xff )
+ {
+ // end of encoded part, check that nothing was left: there can
+ // be up to 4 bits of 0 padding but nothing else (we also need
+ // to check isLSB as we count bits modulo 8 while a valid UTF-7
+ // encoded sequence must contain an integral number of UTF-16
+ // characters)
+ if ( state.isLSB || state.bit > 4 ||
+ (state.accum & ((1 << state.bit) - 1)) )
+ {
+ if ( !len )
+ state = stateOrig;
+
+ return wxCONV_FAILED;
+ }
+
+ state.ToDirect();
+
+ // re-parse this character normally below unless it's '-' which
+ // is consumed by the decoder
+ if ( cc == '-' )
+ continue;
+ }
+ else // valid encoded character
+ {
+ // mini base64 decoder: each character is 6 bits
+ state.bit += 6;
+ state.accum <<= 6;
+ state.accum += dc;
+
+ if ( state.bit >= 8 )
+ {
+ // got the full byte, consume it
+ state.bit -= 8;
+ unsigned char b = (state.accum >> state.bit) & 0x00ff;
+
+ if ( state.isLSB )
+ {
+ // we've got the full word, output it
+ if ( dst )
+ *dst++ = (state.msb << 8) | b;
+ len++;
+ state.isLSB = false;
+ }
+ else // MSB
+ {
+ // just store it while we wait for LSB
+ state.msb = b;
+ state.isLSB = true;
+ }
+ }
+ }
+ }
+
+ if ( state.IsDirect() )
+ {
+ // start of an encoded segment?
+ if ( cc == '+' )
+ {
+ if ( *src == '-' )
+ {
+ // just the encoded plus sign, don't switch to shifted mode
+ if ( dst )
+ *dst++ = '+';
+ len++;
+ src++;
+ }
+ else if ( utf7unb64[(unsigned)*src] == 0xff )
+ {
+ // empty encoded chunks are not allowed
+ if ( !len )
+ state = stateOrig;
+
+ return wxCONV_FAILED;
+ }
+ else // base-64 encoded chunk follows
+ {
+ state.ToShifted();
+ }
+ }
+ else // not '+'
+ {
+ // only printable 7 bit ASCII characters (with the exception of
+ // NUL, TAB, CR and LF) can be used directly
+ if ( cc >= 0x7f || (cc < ' ' &&
+ !(cc == '\0' || cc == '\t' || cc == '\r' || cc == '\n')) )
+ return wxCONV_FAILED;
+
+ if ( dst )
+ *dst++ = cc;
+ len++;
+ }
+ }
+ }
+
+ if ( !len )
+ {
+ // as we didn't read any characters we should be called with the same
+ // data (followed by some more new data) again later so don't save our
+ // state
+ state = stateOrig;
+
+ return wxCONV_FAILED;
+ }
+
+ return len;
+}
+
+//
+// BASE64 encoding table
+//
+static const unsigned char utf7enb64[] =
+{
+ 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H',
+ 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P',
+ 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X',
+ 'Y', 'Z', 'a', 'b', 'c', 'd', 'e', 'f',
+ 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n',
+ 'o', 'p', 'q', 'r', 's', 't', 'u', 'v',
+ 'w', 'x', 'y', 'z', '0', '1', '2', '3',
+ '4', '5', '6', '7', '8', '9', '+', '/'
+};
+
+//
+// UTF-7 encoding table
+//
+// 0 - Set D (directly encoded characters)
+// 1 - Set O (optional direct characters)
+// 2 - whitespace characters (optional)
+// 3 - special characters
+//
+static const unsigned char utf7encode[128] =
+{
+ 0, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 3, 3, 2, 3, 3,
+ 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
+ 2, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 3, 0, 0, 0, 3,
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0,
+ 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 3, 1, 1, 1,
+ 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 3, 3
+};
+
+static inline bool wxIsUTF7Direct(wchar_t wc)
+{
+ return wc < 0x80 && utf7encode[wc] < 1;
+}
+
+size_t wxMBConvUTF7::FromWChar(char *dst, size_t dstLen,
+ const wchar_t *src, size_t srcLen) const
+{
+ EncoderState stateOrig,
+ *statePtr;
+ if ( srcLen == wxNO_LEN )
+ {
+ // we don't apply the stored state when operating on entire strings at
+ // once
+ statePtr = &stateOrig;
+
+ srcLen = wxWcslen(src) + 1;
+ }
+ else // do use the mode we left the output in previously
+ {
+ stateOrig = m_stateEncoder;
+ statePtr = wx_const_cast(EncoderState *, &m_stateEncoder);
+ }
+
+ EncoderState& state = *statePtr;
+
+
+ size_t len = 0;
+
+ const wchar_t * const srcEnd = src + srcLen;
+ while ( src < srcEnd && (!dst || len < dstLen) )
+ {
+ wchar_t cc = *src++;
+ if ( wxIsUTF7Direct(cc) )
+ {
+ if ( state.IsShifted() )
+ {
+ // pad with zeros the last encoded block if necessary
+ if ( state.bit )
+ {
+ if ( dst )
+ *dst++ = utf7enb64[((state.accum % 16) << (6 - state.bit)) % 64];
+ len++;
+ }
+
+ state.ToDirect();
+
+ if ( dst )
+ *dst++ = '-';
+ len++;
+ }
+
+ if ( dst )
+ *dst++ = (char)cc;
+ len++;
+ }
+ else if ( cc == '+' && state.IsDirect() )
+ {
+ if ( dst )
+ {
+ *dst++ = '+';
+ *dst++ = '-';
+ }
+
+ len += 2;
+ }
+#ifndef WC_UTF16
+ else if (((wxUint32)cc) > 0xffff)
+ {
+ // no surrogate pair generation (yet?)
+ return wxCONV_FAILED;
+ }
+#endif
+ else
+ {
+ if ( state.IsDirect() )
+ {
+ state.ToShifted();
+
+ if ( dst )
+ *dst++ = '+';
+ len++;
+ }
+
+ // BASE64 encode string
+ for ( ;; )
+ {
+ for ( unsigned lsb = 0; lsb < 2; lsb++ )
+ {
+ state.accum <<= 8;
+ state.accum += lsb ? cc & 0xff : (cc & 0xff00) >> 8;
+
+ for (state.bit += 8; state.bit >= 6; )
+ {
+ state.bit -= 6;
+ if ( dst )
+ *dst++ = utf7enb64[(state.accum >> state.bit) % 64];
+ len++;
+ }
+ }
+
+ if ( src == srcEnd || wxIsUTF7Direct(cc = *src) )
+ break;
+
+ src++;
+ }
+ }
+ }
+
+ // we need to restore the original encoder state if we were called just to
+ // calculate the amount of space needed as we will presumably be called
+ // again to really convert the data now
+ if ( !dst )
+ state = stateOrig;
+
+ return len;
+}
+
+// ----------------------------------------------------------------------------
+// UTF-8
+// ----------------------------------------------------------------------------
+
+static const 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;
+
+// this table gives the length of the UTF-8 encoding from its first character:
+const unsigned char tableUtf8Lengths[256] = {
+ // single-byte sequences (ASCII):
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 00..0F
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 10..1F
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 20..2F
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 30..3F
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 40..4F
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 50..5F
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 60..6F
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 70..7F
+
+ // these are invalid:
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 80..8F
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 90..9F
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // A0..AF
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // B0..BF
+ 0, 0, // C0,C1
+
+ // two-byte sequences:
+ 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // C2..CF
+ 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // D0..DF
+
+ // three-byte sequences:
+ 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, // E0..EF
+
+ // four-byte sequences:
+ 4, 4, 4, 4, 4, // F0..F4
+
+ // these are invalid again (5- or 6-byte
+ // sequences and sequences for code points
+ // above U+10FFFF, as restricted by RFC 3629):
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 // F5..FF
+};
+
+size_t
+wxMBConvStrictUTF8::ToWChar(wchar_t *dst, size_t dstLen,
+ const char *src, size_t srcLen) const
+{
+ wchar_t *out = dstLen ? dst : NULL;
+ size_t written = 0;
+
+ if ( srcLen == wxNO_LEN )
+ srcLen = strlen(src) + 1;
+
+ for ( const char *p = src; ; p++ )
+ {
+ if ( !(srcLen == wxNO_LEN ? *p : srcLen) )
+ {
+ // all done successfully, just add the trailing NULL if we are not
+ // using explicit length
+ if ( srcLen == wxNO_LEN )
+ {
+ if ( out )
+ {
+ if ( !dstLen )
+ break;
+
+ *out = L'\0';
+ }
+
+ written++;
+ }
+
+ return written;
+ }
+
+ if ( out && !dstLen-- )
+ break;
+
+ wxUint32 code;
+ unsigned char c = *p;
+
+ if ( c < 0x80 )
+ {
+ if ( srcLen == 0 ) // the test works for wxNO_LEN too
+ break;
+
+ if ( srcLen != wxNO_LEN )
+ srcLen--;
+
+ code = c;
+ }
+ else
+ {
+ unsigned len = tableUtf8Lengths[c];
+ if ( !len )
+ break;
+
+ if ( srcLen < len ) // the test works for wxNO_LEN too
+ break;
+
+ if ( srcLen != wxNO_LEN )
+ srcLen -= len;
+
+ // Char. number range | UTF-8 octet sequence
+ // (hexadecimal) | (binary)
+ // ----------------------+----------------------------------------
+ // 0000 0000 - 0000 007F | 0xxxxxxx
+ // 0000 0080 - 0000 07FF | 110xxxxx 10xxxxxx
+ // 0000 0800 - 0000 FFFF | 1110xxxx 10xxxxxx 10xxxxxx
+ // 0001 0000 - 0010 FFFF | 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
+ //
+ // Code point value is stored in bits marked with 'x',
+ // lowest-order bit of the value on the right side in the diagram
+ // above. (from RFC 3629)
+
+ // mask to extract lead byte's value ('x' bits above), by sequence
+ // length:
+ static const unsigned char leadValueMask[] = { 0x7F, 0x1F, 0x0F, 0x07 };
+
+ // mask and value of lead byte's most significant bits, by length:
+ static const unsigned char leadMarkerMask[] = { 0x80, 0xE0, 0xF0, 0xF8 };
+ static const unsigned char leadMarkerVal[] = { 0x00, 0xC0, 0xE0, 0xF0 };
+
+ len--; // it's more convenient to work with 0-based length here
+
+ // extract the lead byte's value bits:
+ if ( (c & leadMarkerMask[len]) != leadMarkerVal[len] )
+ break;
+
+ code = c & leadValueMask[len];
+
+ // all remaining bytes, if any, are handled in the same way
+ // regardless of sequence's length:
+ for ( ; len; --len )
+ {
+ c = *++p;
+ if ( (c & 0xC0) != 0x80 )
+ return wxCONV_FAILED;
+
+ code <<= 6;
+ code |= c & 0x3F;
+ }
+ }
+
+#ifdef WC_UTF16
+ // cast is ok because wchar_t == wxUint16 if WC_UTF16
+ if ( encode_utf16(code, (wxUint16 *)out) == 2 )
+ {
+ if ( out )
+ out++;
+ written++;
+ }
+#else // !WC_UTF16
+ if ( out )
+ *out = code;
+#endif // WC_UTF16/!WC_UTF16
+
+ if ( out )
+ out++;
+
+ written++;
+ }
+
+ return wxCONV_FAILED;
+}
+
+size_t
+wxMBConvStrictUTF8::FromWChar(char *dst, size_t dstLen,
+ const wchar_t *src, size_t srcLen) const
projects / wxWidgets.git / blobdiff