-
Notifications
You must be signed in to change notification settings - Fork 96
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
5 changed files
with
181 additions
and
117 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,118 @@ | ||
Conditional overloading | ||
======================= | ||
|
||
Conditional overloading takes a set of functions and calls the first one that is callable. This is one of the ways to resolve ambiguity with overloads, but more importantly it allows an alternative function to be used when the first is not callable. | ||
|
||
Stringify | ||
--------- | ||
|
||
Take a look at this example of defining a `stringify` function from | ||
stackoverflow [here](http://stackoverflow.com/questions/30189926/metaprograming-failure-of-function-definition-defines-a-separate-function/30515874). | ||
|
||
The user would like to write `stringify` to call `to_string` where applicable | ||
and fallback on using `sstream` to convert to a string. Most of the top | ||
answers usually involve some amount of metaprogramming using either `void_t` | ||
or `is_detected`(see [n4502](http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/n4502.pdf)). However, with the Fit library it can simply be written like | ||
this: | ||
|
||
FIT_STATIC_LAMBDA_FUNCTION(stringify) = conditional( | ||
[](auto x) FIT_RETURNS(to_string(x)), | ||
[](auto x) FIT_RETURNS(static_cast<ostringstream&>(ostringstream() << x).str()) | ||
); | ||
|
||
So, using [`FIT_RETURNS`](returns.md) no only deduces the return type for the function, but it also constrains the function on whether the expression is valid or not either. So by writing `FIT_RETURNS(to_string(x))` then the first function will try to call `to_string` function if possible. If not, then the second function will be called. | ||
|
||
The second function still uses `FIT_RETURNS`, so the function will still be constrained by whether the `<<` stream operator can be used. Although it may seem unnecsarry because there is not another function, however, this makes the function composable. So we could use this to define a `serialize` function that tries to call stringify first, otherwise it looks for the member `.serialize()`: | ||
|
||
FIT_STATIC_LAMBDA_FUNCTION(serialize) = conditional( | ||
[](auto x) FIT_RETURNS(stringify(x)), | ||
[](auto x) FIT_RETURNS(x.serialize()) | ||
); | ||
|
||
static_if | ||
--------- | ||
|
||
In addition, this can be used with the [`fit::if_`](if.md) decorator to create `static_if`-like | ||
constructs. For example, Baptiste Wicht discusses how one could write `static_if` in C++ [here](http://baptiste-wicht.com/posts/2015/07/simulate-static_if-with-c11c14.html). | ||
|
||
He wants to be able to write this: | ||
|
||
template<typename T> | ||
void decrement_kindof(T& value){ | ||
static_if(std::is_same<std::string, T>::value){ | ||
value.pop_back(); | ||
} else { | ||
--value; | ||
} | ||
} | ||
|
||
However, that isn't possible in C++. With the Fit library one can simply write | ||
this: | ||
|
||
template<typename T> | ||
void decrement_kindof(T& value) | ||
{ | ||
eval(conditional( | ||
if_(std::is_same<std::string, T>())([&](auto id){ | ||
id(value).pop_back(); | ||
}), | ||
[&](auto id){ | ||
--id(value); | ||
} | ||
)); | ||
} | ||
|
||
The `id` parameter passed to the lambda is the [`identity`](identity.md) function. As explained in the article, this is used to delay the lookup of types by making it a dependent type(i.e. the type depends on a template parameter), which is necessary to avoid compile errors. | ||
|
||
The advantage of using the Fit library instead of the solution in Baptiste | ||
Wicht's blog, is that [`conditional`](conditional.md) allows more than just two conditions. So if | ||
there was another trait to be checked, such as `is_stack`, it could be written | ||
like this: | ||
|
||
template<typename T> | ||
void decrement_kindof(T& value) | ||
{ | ||
eval(conditional( | ||
if_(is_stack<T>())([&](auto id){ | ||
id(value).pop(); | ||
}), | ||
if_(std::is_same<std::string, T>())([&](auto id){ | ||
id(value).pop_back(); | ||
}), | ||
[&](auto id){ | ||
--id(value); | ||
} | ||
)); | ||
} | ||
|
||
Type traits | ||
----------- | ||
|
||
Furthermore, this technique can be used to write type traits as well. Jens | ||
Weller was looking for a way to define a general purpose detection for pointer | ||
operands(such as `*` and `->`). One way to accomplish this is using Fit like | ||
this: | ||
|
||
// Check that T has member function for operator* and ope | ||
template<class T> | ||
auto has_pointer_member(const T&) -> decltype( | ||
&T::operator*, | ||
&T::operator->, | ||
std::true_type{} | ||
); | ||
|
||
FIT_STATIC_LAMBDA_FUNCTION(has_pointer_operators) = conditional( | ||
FIT_LIFT(has_pointer_member), | ||
[](auto* x) -> bool_constant<(!std::is_void<decltype(*x)>())> { return {}; }, | ||
always(std::false_type{}) | ||
); | ||
|
||
template<class T> | ||
struct is_dereferenceable | ||
: decltype(has_pointer_operators(std::declval<T>())) | ||
{}; | ||
|
||
Which is much simpler than the other implementations that were written, which were | ||
about 3 times the amount of code(see [here](https://gist.github.com/lefticus/6fdccb18084a1a3410d5)). | ||
|
||
The `has_pointer_operators` function works by first trying to call `has_pointer_member` which returns `true_type` if the type has member functions `operator*` and `operator->`, otherwise the function is not callable. The next function is only callable on pointers, which returns true if it is not a `void*` pointer(because `void*` pointers are not dereferenceable). Finally, if none of those functions matched then the last function will always return `false_type`. |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,32 @@ | ||
Polymorphic constructors | ||
======================== | ||
|
||
Writing polymorphic constructors(such as `make_tuple`) is a boilerplate that | ||
has to be written over and over again for template classes: | ||
|
||
template <class T> | ||
struct unwrap_refwrapper | ||
{ | ||
typedef T type; | ||
}; | ||
template <class T> | ||
struct unwrap_refwrapper<std::reference_wrapper<T>> | ||
{ | ||
typedef T& type; | ||
}; | ||
template <class T> | ||
struct unwrap_ref_decay | ||
: unwrap_refwrapper<typename std::decay<T>::type> | ||
{}; | ||
|
||
template <class... Types> | ||
std::tuple<typename unwrap_ref_decay<Types>::type...> make_tuple(Types&&... args) | ||
{ | ||
return std::tuple<typename unwrap_ref_decay<Types>::type...>(std::forward<Types>(args)...); | ||
} | ||
|
||
The [`construct`](construct.md) function takes care of all this boilerplate, and the above can be simply written like this: | ||
|
||
FIT_STATIC_FUNCTION(make_tuple) = construct<std::tuple>().by(decay()); |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,26 @@ | ||
Variadic parameters | ||
=================== | ||
|
||
The library provides utilities to simplify handling variadic parameters. For example, if we want to write a variadic print function that prints each argument, like this: | ||
|
||
print("Hello", "World"); | ||
|
||
We can write this simply by using the [`by`](by.md) adaptor: | ||
|
||
FIT_STATIC_FUNCTION(print) = by(std::cout << _); | ||
|
||
The [`by`](by.md) adaptor calls the function for each argument passed in. In this case, we write `std::cout << _` which uses the [placeholders](placeholders.md) to create a function that print to `std::cout`. | ||
|
||
We can also take binary functions and turn them easily into variadic functions | ||
using [`compress`](compress.md), which will do a fold over the variadic parameters. So a variadic `max` function could be written like | ||
this: | ||
|
||
FIT_STATIC_FUNCTION(max) = compress(FIT_LIFT(std::max)); | ||
|
||
[`FIT_LIFT`](lift.md) is used to grab the entire overload set of `std::max` function, which is needed since `std::max` is templated and we want the variadic `std::max` function to handle any types as well. So now it can be called like this: | ||
|
||
auto n = max(1, 2, 4, 3); // Returns 4 | ||
auto m = max(0.1, 0.2, 0.5, 0.4); // Returns 0.5 | ||
|
||
By using [`compress`](compress.md), `max(1, 2, 4, 3)` is called like `std::max(std::max(std::max(1, 2), 4), 3)` and `max(0.1, 0.2, 0.5, 0.4)` is called like `std::max(std::max(std::max(0.1, 0.2), 0.5), 0.4)`. | ||
|
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters