Substitution failure is not an error: Difference between revisions

Revision as of 07:07, 9 September 2011

Revision as of 15:04, 12 October 2011

Line 8:

<source lang="cpp">

struct Test {

typedef int typefoo;

};

template <typename T>

void f(typename T::typefoo) {} // Definition #1

template <typename T>

Line 19:

int main() {

f<Test>(10); // Call #1.

f<int>(10); // Call #2. Without error (even though there is no int::foo) thanks to SFINAE.

}

</source>

Line 33:

template <typename T>

struct has_typedef_typehas_typedef_foobar {

// Variables "yes" and "no" are guaranteed to have different sizes,

// specifically sizeof(yes) == 1 and sizeof(no) == 2.

Line 40:

template <typename C>

static yes& test(typename C::typefoobar*);

template <typename>

Line 51:

struct foo {

typedef float typefoobar;

};

int main() {

std::cout << std::boolalpha;

std::cout << has_typedef_typehas_typedef_foobar<int>::value << std::endl;

std::cout << has_typedef_typehas_typedef_foobar<foo>::value << std::endl;

}

</source>

When <code>T</code> has the nested type <code>typefoobar</code> defined, the instantiation of the first <code>test</code> works and 0 is successfully passed as the null pointer constant. (And the resulting type of the expression is <code>yes</code>.) If it does not work, the only available function is the second <code>test</code>, and the resulting type of the expression is <code>no</code>. (An ellipsis is used not only because it will accept any argument, but also because its conversion rank is lowest, so a call to the first function will be preferred if it is possible; this removes ambiguity.)

Developers of [[Boost C++ Libraries|Boost]] used SFINAE to great effect in boost::enable_if<ref name="enable_if">[http://www.boost.org/doc/libs/release/libs/utility/enable_if.html Boost Enable If]</ref> and in other ways.