+
+class utf8conv {
+ uint8_t *obfr, *out, *oend;
+ uint8_t *ibfr, *inp, *iend;
+public:
+ utf8conv(void *out, int olen, void *inp, int ilen) {
+ this->obfr = this->out = (uint8_t*)out;
+ this->oend = this->out + olen;
+ this->ibfr = this->inp = (uint8_t*)inp;
+ this->iend = this->inp + ilen;
+ }
+ int cur() { return inp>=iend ? -1 : *inp; }
+ int next() { return inp>=iend ? -1 : *inp++; }
+ int next(int ch) { return out>=oend ? -1 : *out++ = ch; }
+ int ilen() { return inp-ibfr; }
+ int olen() { return out-obfr; }
+ int wnext();
+ int wnext(unsigned int v);
+};
+
+int utf8conv::
+wnext(unsigned int v)
+{
+ if( v < 0x00000080 ) { next(v); return 1; }
+ int n = v < 0x00000800 ? 2 : v < 0x00010000 ? 3 :
+ v < 0x00200000 ? 4 : v < 0x04000000 ? 5 : 6;
+ int m = (0xff00 >> n), i = n-1;
+ next((v>>(6*i)) | m);
+ while( --i >= 0 ) next(((v>>(6*i)) & 0x3f) | 0x80);
+ return n;
+}
+
+int utf8conv::
+wnext()
+{
+ int v = 0, n = 0, ch = next();
+ if( ch == '\\' ) {
+ switch( (ch=next()) ) {
+ case 'n': return '\n';
+ case 't': return '\t';
+ case 'r': return '\r';
+ case 'b': return '\b';
+ case 'f': return '\f';
+ case 'v': return '\v';
+ case 'a': return '\a';
+ case '0': case '1': case '2': case '3':
+ case '4': case '5': case '6': case '7':
+ v = ch - '0';
+ for( int i=3; --i>0; v=v*8+ch, next() )
+ if( (ch=cur()-'0') < 0 || ch >= 8 ) break;
+ return v;
+ case 'x': n = 2; break;
+ case 'u': n = 4; break;
+ case 'U': n = 8; break;
+ default: return ch;
+ }
+ for( int i=n; --i>=0; v=v*16+ch, next() ) {
+ if( (ch=cur()-'0')>=0 && ch<10 ) continue;
+ if( (ch-='A'-'0'-10)>=10 && ch<16 ) continue;
+ if( (ch-='a'-'A')<10 || ch>=16 ) break;
+ }
+ }
+ else if( ch >= 0x80 ) {
+ static const unsigned char byts[] = {
+ 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 4, 5,
+ };
+ int i = ch - 0xc0;
+ n = i<0 ? 0 : byts[i/4];
+ for( v=ch, i=n; --i>=0; v+=next() ) v <<= 6;
+ static const unsigned int ofs[6] = {
+ 0x00000000U, 0x00003080U, 0x000E2080U,
+ 0x03C82080U, 0xFA082080U, 0x82082080U
+ };
+ v -= ofs[n];
+ }
+ else
+ v = ch;
+ return v;
+}
+
+