++p translated weirdly

[kjk]

This example is extracted and simplified from https://github.com/commonmark/cmark/blob/master/src/scanners.c

int _scan_scheme(const unsigned char *p) {
  {
    unsigned char yych;
    yych = *p;
    if (yych <= '@')
      goto yy2;
    if (yych <= 'Z')
      goto yy3;

    yy2:
      ++p;
    yy3 : { return 0; }
    return 3;
  }
}

Is translated to:

// _scan_scheme - transpiled function from  /Users/kjk/src/cmark/test.c:1
func _scan_scheme(p []uint8) (c2goDefaultReturn int) {
    {
        var yych uint8
        yych = p[0]
        if int(yych) <= '@' {
            goto yy2
        }
        if int(yych) <= 'Z' {
            goto yy3
        }
    yy2:
        ;
        p = (*(*[1000000000]uint8)(unsafe.Pointer(uintptr(unsafe.Pointer(&p[0])) + (uintptr)(1)*unsafe.Sizeof(p[0]))))[:]
    yy3:
        ;
        {
            return 0
        }
        return 3
    }
    return
}
func init() {
}

++p; was translated to: p = (*(*[1000000000]uint8)(unsafe.Pointer(uintptr(unsafe.Pointer(&p[0])) + (uintptr)(1)*unsafe.Sizeof(p[0]))))[:]

It seems like the right translation should be p = p[1:].

[Konstantin8105]

Dear @kjk ++p is pointer ariphmetic in C, but in Go we haven't simple access to pointer ariphmetic(we have to use package unsafe), but it is look ugly. Thank you for your solution. Feel free for implemented in c2go.