skip UTF-8 BOM if present

[lemire]

Currently, if you want to parse a string that might have a BOM, you have to do something like this...

std::string input;   
load(input);
input.reserve(input.size() + simdjson::SIMDJSON_PADDING);   
std::string_view json_string(input);   
if (input.size() >= 3 && 0 == memcmp(input.data(), "\xEF\xBB\xBF", 3)) {    
 json_string.remove_prefix(3);  // Skip UTF-8 BOM.   
}   
parser.iterate(json_string, json_string.max_size());

This PR would simplify the code to...

std::string input;   
load(input);
parser.iterate(input);