trouble with partial template specialisations - c++

I have the following class structure
// file foo.h:
struct foo_base
{ ... }
template<typename T> struct foo : foo_base
{ ... };
template<typename F>
using is_foo = std::is_convertible<F,foo_base>;
template<typename, typename=void> struct aux;
template<typename Foo>
struct aux<Foo, typename std::enable_if<is_foo<Foo>::value>::type>
{ ... }; // specialisation for any foo
// file bar.h:
#include "foo.h"
template<typename T> struct bar : foo<T>
{ ... };
template<typename T>
struct aux<bar<T>>
{ ... }; // specialisation for bar<T>
Now, the problem is that for aux<bar<T>> both specialisations provided for aux are viable. Is there a way to avoid this disambiguity without providing another specialisation for every T? Note that modifications to file foo.h must not know about file bar.h.
Note The ambiguity shall be resolved such that the specialisation provided in file bar.h is picked for any aux<bar<T>>. Originally, bar was not a template and the specialisation aux<bar> not partial and hence preferred. The problem arose by making bar a template.

The compiler doesn't see struct aux<bar<T>> as more specialised than struct aux<Foo, typename std::enable_if<is_foo<Foo>::value>::type> because of second template argument. You can specify the second argument the same way in your bar<T> specialisation:
template<typename T>
struct aux<bar<T>, typename std::enable_if<is_foo<bar<T>>::value>::type>
{ };
The rules for how specialised partial template specialisations are are complicated, but I will try to explain very briefly:
The three (your two, plus my one) relevant specialisations are
template<typename Foo>
struct aux<Foo, typename std::enable_if<is_foo<Foo>::value>::type>
template<typename T>
struct aux<bar<T>> // or aux<bar<T>, void>
{ };
template<typename T>
struct aux<bar<T>, typename std::enable_if<is_foo<bar<T>>::value>::type>
{ };
Per the standard (14.5.5.2), to determine which of the class template partial specialisations is the most specialised, the question that needs to be answered is which of the following function template overloads would be the best match in a call to f(aux<bar<T>>()):
template<typename Foo>
void f(aux<Foo, typename std::enable_if<is_foo<Foo>::value>::type>); // 1
template<typename T>
void f(aux<bar<T>>); // or aux<bar<T>, void> // 2
template<typename T>
void f(aux<bar<T>, typename std::enable_if<is_foo<bar<T>>::value>::type>); // 3
And there, in turn, the partial ordering rules for functions say that 1 is not more specialised than 2, and that 2 is not more specialised than 1, roughly speaking, because 1 is not clearly more specialised than 2, and 2 is not clearly more specialised than 1. "Clearly more specialised" is not how the standard words it, but that essentially means that based on the type arguments of one of those, the type arguments of the other are not deducible.
When comparing 1 and 3, however, the arguments of 1 are deducible from 3: Foo can be deduced as bar<T>. Therefore, 3 is at least as specialised as 1. However, the arguments of 3 are not deducible from 1: T cannot be deduced at all. The compiler's conclusion therefore is that 3 is more specialised than 1.

Partial specialization of class templates is based on pattern matching. In contrast, customizing function templates is based on template argument deduction and overload resolution.
Because of the class hierarchy present in your problem, customizing behavior would in principle be more convenient through function template overloading, because that can take derived-to-base conversions into account. The pattern matching used in partial class template specialization does not provide the same flexibility.
However, since C++11, it is possible to do compile-time return type deduction. Here is a solution that combines tag dispatching, default constructors and decltype type deduction:
#include <iostream>
// file foo_base.h:
struct foo_base
{
foo_base() = default;
};
foo_base faux(foo_base const&)
{
return foo_base{};
}
template<class T, class = decltype(faux(T{}))>
struct aux;
template<class T>
struct aux<T, foo_base>
{
enum { value = 1 };
};
// file foo.h:
template<typename T>
struct foo : foo_base
{
foo() = default;
};
// file bar.h:
template<typename T>
struct bar : foo<T>
{
bar() = default;
};
template<class T>
bar<T> faux(bar<T> const&)
{
return bar<T>{};
}
template<class T, class U>
struct aux<T, bar<U>>
{
enum { value = 2 };
};
// file meow.h
template<class T>
struct meow : bar<T>
{
meow() = default;
};
int main()
{
std::cout << aux<foo_base>::value; // 1
std::cout << aux<foo<int>>::value; // 1
std::cout << aux<bar<int>>::value; // 2
std::cout << aux<meow<int>>::value; // 2
}
Live Example that works with both g++ and clang in C++11 mode (C++14 mode is not required!).
The constexpr function faux() is overloaded for foo_bar and as a function template for bar<T>. Any argument whose class is derived from foo_base but not from bar<T> will select the former overload, and anything derived from bar<T> will select the latter overload. This mechanism is the same as e.g. in the Standard Library where iterator categories are used to tag dispatch several implementations of std::advance() e.g.
To use this selection mechanism during partial specialization of your class template aux, two more ingredients are required. First, all classes are required to have a default constructor. Secondly, decltype() is applied to that expression faux(T{}) to deduce the return type.
NOTE: it is not required that faux() is constexpr or that any of the default constructors are constexpr, because decltype() will
not actually evaluate the function call but only deduce its return
type.
The main class template aux has a default template argument:
template<class T, class = decltype(faux(T{}))>
struct aux;
Partially specializing the second argument on foo_base allows you to provide behavior for any class that derives from foo_base:
template<class T>
struct aux<T, foo_base>
{ // custom behavior for anything derived from foo_bar };
The second partial specialization matches any class derived from any template instantiation bar<U>
template<class T, class U>
struct aux<T, bar<U>>
{ // custom behavior for anything derived from bar<U> for some U }
NOTE: the main drawback is that you might need to provide a default constructor to all classes in your hierarchy. This may or may not be an obstacle that you can overcome. Most classes already have a default constructor, but some may not. In that sense this solution is intrusive (i.e. it cannot be bolted on top of existing code, but it requires modification of that code).

Related

Template specialization for the base template type for future derived types

I have a class that works as wrapper for some primitives or custom types. I want to write explicit specialization for custom template type.
My code that reproduces the problem:
template < class T >
struct A {
void func() { std::cout << "base\n"; }
};
template <>
struct A<int> {};
template < class T, class CRTP >
struct BaseCrtp {
void someFunc() {
CRTP::someStaticFunc();
}
};
struct DerrType : BaseCrtp<int, DerrType> {
static void someStaticFunc() {}
};
template < class T, class CRTP >
struct A< BaseCrtp<T, CRTP> > {
void func() { std::cout << "sometype\n"; }
};
int main() {
A<DerrType> a;
a.func(); // print: "base". should be: "sometype"
return 0;
}
A<DerrType> use default function, not a specialization. How can I make specialization for these set of classes?
I will have a lot of types like DerrType, and I want to make common behavior for all of them.
DerrType and others will be used as curiously recurring template pattern
Not sure I fully understood what you want, but maybe something like this:
template<typename T>
concept DerivedFromBaseCrtp = requires(T& t) {
[]<typename U, typename CRTP>(BaseCrtp<U, CRTP>&){}(t);
};
template < DerivedFromBaseCrtp T >
struct A<T> {
void func() { std::cout << "sometype\n"; }
};
The concept basically checks whether T is equal to or is publicly inherited (directly or indirectly) from some specialization of BaseCrtp. Otherwise the call to the lambda would be ill-formed. Template argument deduction only succeeds in the call if the argument and parameter type match exactly or the argument has a derived type of the parameter. If the class is inherited non-publicly, the reference in the call can't bind to the parameter.
The concept will however fail if the type is inherited from multiple BaseCrtp specializations, in which case template argument deduction on the call will not be able to choose between the multiple choices.
Alternatively you can also use the stricter concept
template<typename T>
concept CrtpDerivedFromBaseCrtp = requires(T& t) {
[]<typename U>(BaseCrtp<U, T>&){}(t);
};
which will also require that the type T is actually using the CRTP pattern on BaseCrtp (directly or through a some base class between BaseCrtp and T). Again, this will fail if T is inherited multiple times from some BaseCrtp<U, T> specialization, although it will ignore specializations with a type other than T in the second position.
For another alternative you might want to check that T is derived from some type X such that X is derived from BaseCrtp<U, X> for some U (meaning that X uses the CRTP pattern correctly). That could be done using this variation:
template <typename T>
concept CrtpDerivedFromBaseCrtp =
requires(T& t) {
[]<typename U, typename CRTP>(BaseCrtp<U, CRTP>&)
requires(std::is_base_of_v<CRTP, T> &&
std::is_base_of_v<BaseCrtp<U, CRTP>, CRTP>)
{}
(t);
};
Again, this fails if T is derived from multiple BaseCrtp specializations, directly or indirectly.

How to conditionally get template type of multiple base class with multiple inheritance

This is NOT a duplicate of link
Consider the following code:
#include <type_traits>
template <typename... Bases>
struct Overloads : public Bases... {};
template <typename T>
struct A {
using AType = T;
};
template <typename T>
struct B {
using BType = T;
};
template <typename T>
struct C {
using CType = T;
};
template <typename OverloadsType>
struct Derived : public OverloadsType {
};
int main() {
// OK
static_assert(std::is_same_v<typename Derived<Overloads<A<int>, B<float>, C<char>>>::AType, int>);
// OK
static_assert(std::is_same_v<typename Derived<Overloads<A<int>, B<float>, C<char>>>::BType, float>);
// OK
static_assert(std::is_same_v<typename Derived<Overloads<A<int>, B<float>, C<char>>>::CType, char>);
// ???
static_assert(std::is_same_v<typename Derived<Overloads<B<float>, C<char>>>::AType, void>);
}
Demo: Link
For the last line, I need to detect that Derived<Overloads<B<float>, C<char>>> is NOT derived from A, so I want typename Derived<Overloads<B<float>, C<char>>>::AType to be void or something (it fails to compile)
How can I do this?
My precise use case is: 1. Determine if Derived is derived from A<T> for some T. 2. If so, figure out that T.
With C++20 concepts, this isn't too difficult. You need a function that takes A<T> as a parameter. There need not be a function definition; we're just using template argument deduction. It will never be called:
template<typename T>
T getDerivedFromAType(A<T> const&); //Not implemented, since we will never call it.
Any type U for which getDerivedFromAType(u) works will either be an A<T> itself or a type derived from A<T> (from a single A<T> of course). So we can build a concept out of it:
template<typename U>
concept IsDerivedFromA = requires(U u)
{
getDerivedFromAType(u);
};
So long as nobody writes an overload of getDerivedFromAType, you're fine. This function ought to be in a detail namespace or have something else that lets people know that it is off-limits.
To get the actual T used by A... well, there's a reason why the function returned T. We simply need a using statement that calculates the return type of the function call:
template<IsDerivedFromA T>
using BaseAType = decltype(getDerivedFromAType(std::declval<T>()));
Notice that the template is guarded by our concept.

Divorce a parameter pack in a class template

I am trying to write a class template that uses a parameter-pack and implements a member function for each type contained in the parameter-pack.
This is what I have so far:
template <typename...T>
class Myclass {
public:
void doSomething((Some_Operator_to_divorce?) T) {
/*
* Do Something
*/
std::cout << "I did something" << std::endl;
}
};
My goal is to have a class template that can be used in the following way:
Myclass<std::string, int, double> M;
M.doSomething("I am a String");
M.doSomething(1234);
M.doSomething(0.1234);
Where the class template mechanism will create an implementation for a doSomething(std::string x), a doSomething(int x) and a doSomething(double x) member function but not a doSomething(std::string x, int i, double f) member function.
I found a lot of examples in the web on the usability of parameter-packs, but I could not figure out if it can be used for my purpose, or if I totally misunderstood for what a parameter-pack can be used.
I thought that I need to unpack the parameter-pack but, after reading a lot of examples about unpacking parameter packs, I believe that this is not the right choice and it has a complete different meaning.
So, therefore, I am looking for a operation to "divorce" a parameter-pack.
There is no "operator" specifically that supports this, but what you're requesting can be done in a few different ways, depending on your requirements.
The only way to "extract" T types from a parameter pack of a class template with the purpose of implementing an overload-set of functions is to implement it using recursive inheritance, where each instance extracts one "T" type and implements the function, passing the rest on to the next implementation.
Something like:
// Extract first 'T', pass on 'Rest' to next type
template <typename T, typename...Rest>
class MyClassImpl : public MyClassImpl<Rest...>
{
public:
void doSomething(const T&) { ... }
using MyClassImpl<Rest...>::doSomething;
};
template <typename T>
class MyClassImpl<T> // end-case, no more 'Rest'
{
public:
void doSomething(const T&) { ... }
};
template <typename...Types>
class MyClass : public MyClassImpl<Types...>
{
public:
using MyClassImpl<Types...>::doSomething;
...
};
This will instantiate sizeof...(Types) class templates, where each one defines an overload for each T type.
This ensures that you get overload semantics -- such that passing an int can call a long overload, or will be ambiguous if there are two competing conversions.
However, if this is not necessary, then it'd be easier to enable the function with SFINAE using enable_if and a condition.
For exact comparisons, you could create an is_one_of trait that only ensures this exists if T is exactly one of the types. In C++17, this could be done with std::disjunction and std::is_same:
#include <type_traits>
// A trait to check that T is one of 'Types...'
template <typename T, typename...Types>
struct is_one_of : std::disjunction<std::is_same<T,Types>...>{};
Alternatively, you may want this to only work if it may work with convertible types -- which you might do something like:
template <typename T, typename...Types>
struct is_convertible_to_one_of : std::disjunction<std::is_convertible<T,Types>...>{};
The difference between the two is that if you passed a string literal to a MyClass<std::string>, it will work with the second option since it's convertible, but not the first option since it's exact. The deduced T type from the template will also be different, with the former being exactly one of Types..., and the latter being convertible (again, T may be const char*, but Types... may only contain std::string)
To work this together into your MyClass template, you just need to enable the condition with SFINAE using enable_if:
template <typename...Types>
class MyClass
{
public:
// only instantiates if 'T' is exactly one of 'Types...'
template <typename T, typename = std::enable_if_t<is_one_of<T, Types...>::value>>
void doSomething(const T&) { ... }
// or
// only instantiate if T is convertible to one of 'Types...'
template <typename T, typename = std::enable_if_t<is_convertible_to_one_of<T, Types...>::value>>
void doSomething(const T&) { ... }
};
Which solution works for you depends entirely on your requirements (overload semantics, exact calling convension, or conversion calling convension)
Edit: if you really wanted to get complex, you can also merge the two approaches... Make a type trait to determine what type would be called from an overload, and use this to construct a function template of a specific underlying type.
This is similar to how variant needs to be implemented, since it has a U constructor that considers all types as an overload set:
// create an overload set of all functions, and return a unique index for
// each return type
template <std::size_t I, typename...Types>
struct overload_set_impl;
template <std::size_t I, typename T0, typename...Types>
struct overload_set_impl<I,T0,Types...>
: overload_set_impl<I+1,Types...>
{
using overload_set_impl<I+1,Types...>::operator();
std::integral_constant<std::size_t,I> operator()(T0);
};
template <typename...Types>
struct overload_set : overload_set_impl<0,Types...> {};
// get the index that would be returned from invoking all overloads with a T
template <typename T, typename...Types>
struct index_of_overload : decltype(std::declval<overload_set<Types...>>()(std::declval<T>())){};
// Get the element from the above test
template <typename T, typename...Types>
struct constructible_overload
: std::tuple_element<index_of_overload<T, Types...>::value, std::tuple<Types...>>{};
template <typename T, typename...Types>
using constructible_overload_t
= typename constructible_overload<T, Types...>::type;
And then use this with the second approach of having a function template:
template <typename...Types>
class MyClass {
public:
// still accept any type that is convertible
template <typename T, typename = std::enable_if_t<is_convertible_to_one_of<T, Types...>::value>>
void doSomething(const T& v)
{
// converts to the specific overloaded type, and call it
using type = constructible_overload_t<T, Types...>;
doSomethingImpl<type>(v);
}
private:
template <typename T>
void doSomethingImpl(const T&) { ... }
This last approach does it two-phase; it uses the first SFINAE condition to ensure it can be converted, and then determines the appropriate type to treat it as and delegates it to the real (private) implementation.
This is much more complex, but can achieve the overload-like semantics without actually requiring recursive implementation in the type creating it.

Template Specialisation with Template Argument

Let's suppose to have a templateclass Foo:
template <typename T>
class Foo {
void foo();
};
I have another template class Bar (independent from the first one):
template <int N>
class Bar {};
Let's say, I want to specialise the foo() method for whatever Bar class.
I'd wrongly write:
template <>
template <int N>
void Foo<Bar<N> >::foo() { /* ... */ }
The compiler blames me for because the type is not complete:
error: invalid use of incomplete type 'class Foo<Bar<N> >'
void Foo<Bar<N> >::foo() { }
Code
I am using C++98, but I'd like to know if there exist different solutions in C++11.
Note
I could solve the problem specialising the entire class Foo for a generic Bar, but after I should have to define all methods.
Example Code
That's not what I want, I am looking for (if exists) more elegant solution (both C++98 and C++11) which allows me to specialise and implement only a single class method.
EDIT:
The question on SO does not explain how to specialise with a template argument. Indeed, my question shows how the compiler complains about that.
For C++11 you can SFINAE enable/disable (using std::enable_if) two differents versions of foo() inside a not specialized Foo class.
In C++98 you don't have std::enable_if but you can simulate it (give me some minutes and I try to propose an example). Sorry: my idea doesn't works because this method require the use of default template arguments for methods that is a C++11 innovation.
Another way is define a template base class for Foo(), say FooBase, insert foo() (and only foo()) in FooBase and specialize FooBase.
Another way, that works also with C++98, can be tag dispatching: you can define an unique foo(), with zero parameter, that call another foo(), with a parameter that is determined by T.
The following is a full (C++98 compilable) example
#include <iostream>
struct barWay {};
struct noBarWay {};
template <int>
struct Bar
{ };
template <typename>
struct selectType
{ typedef noBarWay type; };
template <int N>
struct selectType< Bar<N> >
{ typedef barWay type; };
template <typename T>
struct Foo
{
void foo (noBarWay const &)
{ std::cout << "not Bar version" << std::endl; }
void foo (barWay const &)
{ std::cout << "Bar version" << std::endl; }
void foo ()
{ foo(typename selectType<T>::type()); }
};
int main ()
{
Foo<int> fi;
Foo< Bar<42> > fb;
fi.foo();
fb.foo();
}
if a common base is not desirable, yet another way could be giving foo() a customization point, like a trait for example:
template <typename T>
struct foo_traits;
template <typename T>
struct Foo {
void foo(){ foo_traits<T>::foo_cp(*this); }
};
template <typename T>
struct foo_traits{ static void foo_cp(T&){/*default*/} };
template <int N>
class Bar {};
template <int N>
struct foo_traits<Bar<N>>{ static void foo_cp(Foo<Bar<N>>&){/*spec*/} };
such trait could also be an implementation detail friend, if its only purpose is to internally provide a foo() specialization for Bar's.
If you cannot specialize foo, define it so that it delegates the call to an internal foo-implementation class. Then specialize that class.
Something like this should compile in C++98 and it doesn't differ much from your original code:
template <typename T>
class Foo {
template<typename>
struct FooImpl;
public:
void foo() { FooImpl<T>()(); }
};
template <int N>
class Bar {};
template <typename T>
template <int N>
struct Foo<T>::FooImpl< Bar<N> > {
void operator()() { /* ... */ }
};
int main() {
Foo< Bar<0> > fb;
fb.foo();
Foo<int> fi;
//fi.foo();
}
The last line doesn't compile as expected (at least I got it was the expected result, just define the function call operator for FooImpl otherwise).
This way you can define selectively the specializations for which you want foo to work. In all the other cases, an attempt at using foo will result in a compilation error.
I'd like to know if there exist different solutions in C++11.
This is a classic use case for tagged dispatch, of which max66 already suggested. The approach, and even syntax, are basically the same in C++98 and C++11.
Here's a bit of a cleaner implementation than max66's, I believe (running on godbolt):
template <class T>
class Foo {
template <class>
struct tag{};
template<class U>
void foo_helper(tag<U>){std::cout << "default\n";}
void foo_helper(tag<Bar<3> >){std::cout << "specialization for Bar<3>\n";}
public:
void foo(){return foo_helper(tag<T>());}
};
The principle is the same; a client function accepting no arguments calls a helper function that constructs an empty type based on the T argument. Then normal overloading takes care of the rest.
Only here I use a templated catch-all method.
In C++11 the syntax would only change slightly; We could say tag<Bar<3>> instead of tag<Bar<3> > because new parsing rules allow the chevron for nested templates.
We could also make the tag and the templated foo_helper catch-all into variadic templates to be a little more generic:
template <class T>
class Foo {
template <class...>
struct tag{};
template<class... U>
void foo_helper(tag<U...>){std::cout << "default\n";}
void foo_helper(tag<Bar<3>>){std::cout << "specialization for Bar<3>\n";}
public:
void foo(){return foo_helper(tag<T>{});}
};
Things actually start getting pretty interesting in C++17 with the introduction of constexpr if that allows us to write what looks like normal branching logic based on T (Live Demo):
template <class T>
class Foo {
public:
void foo(){
if constexpr (std::is_same_v<T, Bar<3>>){std::cout << "Specialization for Bar<3>\n";}
else std::cout << "default\n";
}
};
As you can see, all the tag stuff goes away in favor of using a simple if statement.
We take advantage of type_traits introduced in C++11 to check the type of T against our desired type. Something like this wouldn't necessarily work previously because all branches needed to be compiled. In C++17, only the branch that is selected (at compile-time) is compiled.
Note that you could emulate this behavior as early as C++98 by using typeid (godbolt demo):
void foo(){
if (typeid(T) == typeid(Bar<3>)){std::cout << "Specialization for Bar<3>\n";}
else std::cout << "default\n";
}
However, the typeid approach is a poor choice for 2 reasons:
It's a run time check (slow) for information we know at compile-time
It's brittle because all branches must compile for all template instantiations, whereas in C++17 if constexpr only compiles the branch that is selected.

signed/unsigned trait programming

I'm starting to learn about traits and templates in c++. What I'm wondering is is it possible to create templates for signed/unsigned integral types. The idea is that the normal class would (probably) be implemented for singed integer types, and the variation for unsigned integer types. I tried:
template <typename T>
class FXP<T>
{ ... };
template <typename T>
class FXP<unsigned T>
{ ... };
but this does not compile.
I even came across:
std::is_integral
std::is_signed
std::is_unsigned
So, how do I put these in action to define a class that only supports these two variants?
In this case, there's a few ways to go about it, but my favorite for cases where there's a limited number of variants (e.g., a boolean or two saying which way it should behave), partial specialization of a template is usually the best way to go:
// Original implementation with default boolean for the variation type
template <typename T, bool is_unsigned = std::is_unsigned<T>::value>
class FXP {
// default implementation here
};
Your next step is to then provide a partial specialization that takes the typename T, but only works for a specific variant of the template parameters (e.g. true or false).
template <typename T>
class FXP<T, false> {
// partial specialization when is_unsigned becomes false
};
template <typename T>
class FXP<T, true> {
// partial specialization when is_unsigned becomes true
};
In this case, if you write a default implementation, you only need to make a specialization for the case that's non-default (like the true case).
Here's an example, where the default case gets overridden by a specialized template parameter:
http://coliru.stacked-crooked.com/a/bc761b7b44b0d452
Note that this is better only for smaller cases. If you need complex tests, you're better off using std::enable_if and some more complicated template parameters (like in DyP's answer).
Good luck!
With an additional template parameter:
#include <iostream>
#include <type_traits>
template <typename T, class X = void>
struct FXP
{
// possibly disallow using this primary template:
// static_assert(not std::is_same<X, X>{},
// "Error: type neither signed nor unsigned");
void print() { std::cout << "non-specialized\n"; }
};
template <typename T>
struct FXP< T, typename std::enable_if<std::is_signed<T>{}>::type >
{ void print() { std::cout << "signed\n"; } };
template <typename T>
struct FXP< T, typename std::enable_if<std::is_unsigned<T>{}>::type >
{ void print() { std::cout << "unsigned\n"; } };
struct foo {};
int main()
{
FXP<foo>().print();
FXP<int>().print();
FXP<unsigned int>().print();
}