At work I recently wrote a less than operator for a class implemented from a published specification and which has many properties, six of which are used to uniquely identify an instance of the class. (For the sake of this question, we'll call these properties a-f.) Also, these six properties are of six different types. I defined the operator as such:
bool operator<(const Class& lhs, const Class& rhs)
{
bool retval = (&lhs != &rhs);
if (retval == true)
{
if (lhs.a == rhs.a)
{
if (lhs.b == rhs.b)
{
if (lhs.c == rhs.c)
{
if (lhs.d == rhs.d)
{
if (lhs.e == rhs.e)
{
retval = (lhs.f < rhs.f);
} else {
retval = (lhs.e < rhs.e);
}
} else {
retval = (lhs.d < rhs.d);
}
} else {
retval = (lhs.c < rhs.c);
}
} else {
retval = (lhs.b < rhs.b);
}
} else {
retval = (lhs.a < rhs.a);
}
}
return retval;
}
This, of course, breaks the Linux kernel coding philosophy of, "If you need more than 3 levels of indentation, you're screwed anyway, and should fix your program." So my question is, is there a better way to define this operator to not have so many levels of indentation?
You can write this kind of lexicographical comparison like this:
if (lhs.a != rhs.a) return lhs.a < rhs.a;
if (lhs.b != rhs.b) return lhs.b < rhs.b;
if (lhs.c != rhs.c) return lhs.c < rhs.c;
if (lhs.d != rhs.d) return lhs.d < rhs.d;
if (lhs.e != rhs.e) return lhs.e < rhs.e;
return lhs.f < rhs.f;
You can re-write this with a single return like this:
bool result;
if (lhs.a != rhs.a) result = lhs.a < rhs.a;
else if (lhs.b != rhs.b) result = lhs.b < rhs.b;
else if (lhs.c != rhs.c) result = lhs.c < rhs.c;
else if (lhs.d != rhs.d) result = lhs.d < rhs.d;
else if (lhs.e != rhs.e) result = lhs.e < rhs.e;
else result = lhs.f < rhs.f;
return result;
You can use std::tie
to do lexicographical comparisons:
bool operator<(const Class& lhs, const Class& r) {
return std::tie(lhs.a, lhs.b, lhs.c, lhs.d, lhs.e) < std::tie(rhs.a, rhs.b, rhs.c, rhs.d, rhs.e);
}
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With