Incorrect == for decimal64

[libbooze]

Following code:

void compare_dec64()
{
    using namespace boost::decimal;
    using Dec= decimal64;
    std::print("{}\n", boost::typeindex::type_id<Dec>().pretty_name());
    const Dec a{1452241920, 1};
    const Dec b{ 9223372037, 11};
    std::print("max:{:.40e}\n", std::numeric_limits<Dec>::max());
    std::print("a  : {:.40e}\n", a);
    std::print("b  : {:.40e}\n", b);
    std::print("same {} \n\n", a  == b );
}

prints(on branch limits since master does not print max correctly):

boost::decimal::decimal64
max:9.9999999999999990000000000000000000000000e+384
a  : 1.4522419200000000000000000000000000000000e+10
b  : 9.2233720370000000000000000000000000000000e+20
same true 

I am pretty sure issue is uint64 overflow in

    delta_exp >= 0 ? lhs_sig *= detail::pow10(static_cast<comp_type>(delta_exp)) :
                     rhs_sig *= detail::pow10(static_cast<comp_type>(-delta_exp));

While I believe first intuition for fix will bet to just widen the operands I am actually not sure if that is best fix.

Not sure if this is limited to decimal64, took me a while to get correct values to trigger a bug, did not have time to check for other types.