The Amazing C UTF-8 validator/XML escaper

gistfile1.c

#adapted from http://stackoverflow.com/a/2977559

#define IS_IN_RANGE(c, f, l)    (((c) >= (f)) && ((c) <= (l)))
int validUTF8(char *text, int *data_offset);

int validUTF8(char *text, int *data_offset)
{
    if(data_offset)
        *data_offset = 0;
    
    unsigned char b, b2;
    unsigned char *ptr = (unsigned char*) text;
    
    int i = 0;
    int seqlen;
    
    while( ptr[i] != 0 ) {
        b = ptr[i];
        
        if( (b & 0x80) == 0 ) {
            seqlen = 1;
        } else if((b & 0xE0) == 0xC0) {
            seqlen = 2;
        } else if((b & 0xF0) == 0xE0) {
            seqlen = 3;
        } else if((b & 0xF8) == 0xF0) {
            seqlen = 4;
        } else {
            return -1;
        }
        
        for (int j = 1; j < seqlen; ++j) {
            b = ptr[i+j];
            if((b & 0xC0) != 0x80)
                return -1;
        }
        
        switch (seqlen) {
            case 2:
            {
                b = ptr[i];
                if(!IS_IN_RANGE(b, 0xC2, 0xDF)) {
                    return -1;
                }
                break;
            }
                
            case 3:
            {
                b = ptr[i];
                b2 = ptr[i+1];
                
                if(((b == 0xE0) && !IS_IN_RANGE(b2, 0xA0, 0xBF)) ||
                   ((b == 0xED) && !IS_IN_RANGE(b2, 0x80, 0x9F)) ||
                   (!IS_IN_RANGE(b, 0xE1, 0xEC) && !IS_IN_RANGE(b, 0xEE, 0xEF)) ) {
                    return -1;
                }
                break;
            }
                
            case 4:
            {
                b = ptr[i];
                b2 = ptr[i+1];
                
                if(((b == 0xF0) && !IS_IN_RANGE(b2, 0x90, 0xBF)) ||
                   ((b == 0xF4) && !IS_IN_RANGE(b2, 0x80, 0x8F)) ||
                   !IS_IN_RANGE(b, 0xF1, 0xF3)) {
                    return -1;
                }
                break;
            }
        }
        i += seqlen;
    }
    
    if( data_offset )
        *data_offset = i;
    
    return 1;
}

char *input_text;
int textoffset;
int i = 0;
while(input_text[i] != 0) {
    if( validUTF8(&input_text[i], &textoffset) == 1 )
    {
      for (int j=0; j < textoffset; j++) {
         printf ("%s",&input_text[i]);
         i++;
      }
    }
    else
    {
            //assume win-1252 encoding
            int codepoint = input_text[i] + 256;
            if (IS_IN_RANGE(codepoint, 128, 159)) {
                switch (codepoint) {
                    case 128: {codepoint = 0x20AC; break;}
                    case 130: {codepoint = 0x201A; break;}
                    case 131: {codepoint = 0x0192; break;}
                    case 132: {codepoint = 0x201E; break;}
                    case 133: {codepoint = 0x2026; break;}
                    case 134: {codepoint = 0x2020; break;}
                    case 135: {codepoint = 0x2021; break;}
                    case 136: {codepoint = 0x02C6; break;}
                    case 137: {codepoint = 0x2030; break;}
                    case 138: {codepoint = 0x0160; break;}
                    case 139: {codepoint = 0x2039; break;}
                    case 140: {codepoint = 0x0152; break;}
                    case 142: {codepoint = 0x017D; break;}
                    case 145: {codepoint = 0x2018; break;}
                    case 146: {codepoint = 0x2019; break;}
                    case 147: {codepoint = 0x201C; break;}
                    case 148: {codepoint = 0x201D; break;}
                    case 149: {codepoint = 0x2022; break;}
                    case 150: {codepoint = 0x2013; break;}
                    case 151: {codepoint = 0x2014; break;}
                    case 152: {codepoint = 0x02DC; break;}
                    case 153: {codepoint = 0x2122; break;}
                    case 154: {codepoint = 0x0161; break;}
                    case 155: {codepoint = 0x203A; break;}
                    case 156: {codepoint = 0x0153; break;}
                    case 158: {codepoint = 0x017E; break;}
                    case 159: {codepoint = 0x0178; break;}
                }
            }
            printf ("&#%d;", codepoint);
            i++;
        }
    }