C++ function template partial specialization?

Question

I know that the below code is a partial specialization of a class:

template <typename T1, typename T2> 
class MyClass { 
  … 
}; 


// partial specialization: both template parameters have same type 
template <typename T> 
class MyClass<T,T> { 
  … 
}; 

Also I know that C++ does not allow function template partial specialization (only full is allowed). But does my code mean that I have partially specialized my function template for one/same type arguments? Because it works for Microsoft Visual Studio 2010 Express! If no, then could you please explain the partial specialization concept?

#include <iostream>
using std::cin;
using std::cout;
using std::endl;

template <typename T1, typename T2> 
inline T1 max (T1 const& a, T2 const& b) 
{ 
    return a < b ? b : a; 
} 

template <typename T> 
inline T const& max (T const& a, T const& b)
{
    return 10;
}


int main ()
{
    cout << max(4,4.2) << endl;
    cout << max(5,5) << endl;
    int z;
    cin>>z;
}

Answer 1

Function partial specialization is not yet allowed as per the standard. In the example, you are actually overloading & not specializing the max<T1,T2> function.
Its syntax should have looked somewhat like below, had it been allowed:

// Partial specialization is not allowed by the spec, though!
template <typename T> 
inline T const& max<T,T> (T const& a, T const& b)
{                  ^^^^^ <--- [supposed] specializing here
  return 10;
}

In the case of a function templates, only full specialization is allowed by the C++ standard, -- excluding the compiler extensions!

Answer 2

Since partial specialization is not allowed -- as other answers pointed --, you could work around it using std::is_same and std::enable_if, as below:

template <typename T, class F>
inline typename std::enable_if<std::is_same<T, int>::value, void>::type
typed_foo(const F& f) {
    std::cout << ">>> messing with ints! " << f << std::endl;
}

template <typename T, class F>
inline typename std::enable_if<std::is_same<T, float>::value, void>::type
typed_foo(const F& f) {
    std::cout << ">>> messing with floats! " << f << std::endl;
}

int main(int argc, char *argv[]) {
    typed_foo<int>("works");
    typed_foo<float>(2);
}

Output:

$ ./a.out 
>>> messing with ints! works
>>> messing with floats! 2

---

Edit: In case you need to be able to treat all the other cases left, you could add a definition which states that already treated cases should not match -- otherwise you'd fall into ambiguous definitions. The definition could be:

template <typename T, class F>
inline typename std::enable_if<(not std::is_same<T, int>::value)
    and (not std::is_same<T, float>::value), void>::type
typed_foo(const F& f) {
    std::cout << ">>> messing with unknown stuff! " << f << std::endl;
}

int main(int argc, char *argv[]) {
    typed_foo<int>("works");
    typed_foo<float>(2);
    typed_foo<std::string>("either");
}

Which produces:

$ ./a.out 
>>> messing with ints! works
>>> messing with floats! 2
>>> messing with unknown stuff! either

Although this all-cases thing looks a bit boring, since you have to tell the compiler everything you've already done, it's quite doable to treat up to 5 or a few more specializations.

Answer 3

What is specialization ?

If you really want to understand templates, you should take a look at functional languages. The world of templates in C++ is a purely functional sublanguage of its own.

In functional languages, selections are done using Pattern Matching:

-- An instance of Maybe is either nothing (None) or something (Just a)
-- where a is any type
data Maybe a = None | Just a

-- declare function isJust, which takes a Maybe
-- and checks whether it's None or Just
isJust :: Maybe a -> Bool

-- definition: two cases (_ is a wildcard)
isJust None = False
isJust Just _ = True

As you can see, we overload the definition of isJust.

Well, C++ class templates work exactly the same way. You provide a main declaration, that states the number and nature of the parameters. It can be just a declaration, or also acts as a definition (your choice), and then you can (if you so wish) provide specializations of the pattern and associate to them a different (otherwise it would be silly) version of the class.

For template functions, specialization is somewhat more awkward: it conflicts somewhat with overload resolution. As such, it has been decided that a specialization would relate to a non-specialized version, and specializations would not be considered during overload resolution. Therefore, the algorithm for selecting the right function becomes:

1. Perform overload resolution, among regular functions and non-specialized templates
2. If a non-specialized template is selected, check if a specialization exist for it that would be a better match

(for on in-depth treatment, see GotW #49)

As such, template specialization of functions is a second-zone citizen (literally). As far as I am concerned, we would be better off without them: I have yet to encounter a case where a template specialization use could not be solved with overloading instead.

Is this a template specialization ?

No, it is simply an overload, and this is fine. In fact, overloads usually work as we expect them to, while specializations can be surprising (remember the GotW article I linked).

Answer 4

Non-class, non-variable partial specialization is not allowed, but as said:

All problems in computer science can be solved by another level of indirection. —— David Wheeler

Adding a class to forward the function call can solve this, here is an example:

template <class Tag, class R, class... Ts>
struct enable_fun_partial_spec;

struct fun_tag {};

template <class R, class... Ts>
constexpr R fun(Ts&&... ts) {
  return enable_fun_partial_spec<fun_tag, R, Ts...>::call(
      std::forward<Ts>(ts)...);
}

template <class R, class... Ts>
struct enable_fun_partial_spec<fun_tag, R, Ts...> {
  constexpr static R call(Ts&&... ts) { return {0}; }
};

template <class R, class T>
struct enable_fun_partial_spec<fun_tag, R, T, T> {
  constexpr static R call(T, T) { return {1}; }
};

template <class R>
struct enable_fun_partial_spec<fun_tag, R, int, int> {
  constexpr static R call(int, int) { return {2}; }
};

template <class R>
struct enable_fun_partial_spec<fun_tag, R, int, char> {
  constexpr static R call(int, char) { return {3}; }
};

template <class R, class T2>
struct enable_fun_partial_spec<fun_tag, R, char, T2> {
  constexpr static R call(char, T2) { return {4}; }
};

static_assert(std::is_same_v<decltype(fun<int>(1, 1)), int>, "");
static_assert(fun<int>(1, 1) == 2, "");

static_assert(std::is_same_v<decltype(fun<char>(1, 1)), char>, "");
static_assert(fun<char>(1, 1) == 2, "");

static_assert(std::is_same_v<decltype(fun<long>(1L, 1L)), long>, "");
static_assert(fun<long>(1L, 1L) == 1, "");

static_assert(std::is_same_v<decltype(fun<double>(1L, 1L)), double>, "");
static_assert(fun<double>(1L, 1L) == 1, "");

static_assert(std::is_same_v<decltype(fun<int>(1u, 1)), int>, "");
static_assert(fun<int>(1u, 1) == 0, "");

static_assert(std::is_same_v<decltype(fun<char>(1, 'c')), char>, "");
static_assert(fun<char>(1, 'c') == 3, "");

static_assert(std::is_same_v<decltype(fun<unsigned>('c', 1)), unsigned>, "");
static_assert(fun<unsigned>('c', 1) == 4, "");

static_assert(std::is_same_v<decltype(fun<unsigned>(10.0, 1)), unsigned>, "");
static_assert(fun<unsigned>(10.0, 1) == 0, "");

static_assert(
    std::is_same_v<decltype(fun<double>(1, 2, 3, 'a', "bbb")), double>, "");
static_assert(fun<double>(1, 2, 3, 'a', "bbb") == 0, "");

static_assert(std::is_same_v<decltype(fun<unsigned>()), unsigned>, "");
static_assert(fun<unsigned>() == 0, "");

Answer 5

No. For example, you can legally specialize std::swap, but you cannot legally define your own overload. That means that you cannot make std::swap work for your own custom class template.

Overloading and partial specialization can have the same effect in some cases, but far from all.

Answer 6

Late answer, but some late readers might find it useful: Sometimes, a helper function – designed such that it can be specialised – can solve the issue, too.

So let's imagine, this is what we tried to solve:

template <typename R, typename X, typename Y>
void function(X x, Y y)
{
    R* r = new R(x);
    f(r, y); // another template function?
}

// for some reason, we NEED the specialization:
template <typename R, typename Y>
void function<R, int, Y>(int x, Y y) 
{
    // unfortunately, Wrapper has no constructor accepting int:
    Wrapper* w = new Wrapper();
    w->setValue(x);
    f(w, y);
}

OK, partial template function specialisation, we cannot do that... So let's "export" the part needed for specialisation into a helper function, specialize that one and use it:

template <typename R, typename T>
R* create(T t)
{
    return new R(t);
}
template <>
Wrapper* create<Wrapper, int>(int n) // fully specialized now -> legal...
{
    Wrapper* w = new Wrapper();
    w->setValue(n);
    return w;
}

template <typename R, typename X, typename Y>
void function(X x, Y y)
{
    R* r = create<R>(x);
    f(r, y); // another template function?
}

This can be interesting especially if the alternatives (normal overloads instead of specialisations, the workaround proposed by Rubens, ... – not that these are bad or mine is better, just another one) would share quite a lot of common code.