CODEWITHSUNDEEP
c++11
Andrei Nemes
Andrei Nemes

Reputation: 3122

Use case of constexpr versus defines

I've been using C++ for a bit but never really got into great detail. Recently found out about constexpr and I think it's got great potential for micro-optimizations, though I struggle to see how to properly use it. Considering the following:

#define squared(n) (n * n)

and

constexpr int squared(int n) {
    return n * n;
}

They both output the same thing, but I'm sure they are very different. From my knowledge define directives get replaced at compile-time so any instance of squared(n) simply becomes (n * n).

Obviously this is a simple example, but I'd like to know where does constexpr come into play? Is it meant to used for operations more complex than simple text substitution? What is the main advantage of using it and how is it different from normal expressions?

Upvotes: 3

Views: 707

Answers (1)

Angew is no longer proud of SO
Angew is no longer proud of SO

Reputation: 171167

They don't output the same thing at all. Consider this call:

squared(x++);

Or this class:

struct Number
{
  float value() const;
  float squared() const;
  // ... other members
};

Both of these will fail miserably if squared is a function-like macro.

So, as always: don't use macros where another feature of the language can work as well. You could say that constexpr simply allows us to use macros in fewer places - a good thing.

And there's one more, very important distinction: constexpr functions can be (and indeed often are) recursive. Function-like macros cannot.

Upvotes: 5

Related Questions