This repository contains a proposal to add any_iterator and some helper alias templates to the C++ standard library.
The proposal paper can be found as "any_iterator.md" in this respository (link here).
Yes, there is a sample implementation of that proposal within the "implementation" folder.
To set-up and build the sample implementation and tests, perform the following steps:
git clone [email protected]:Shaktal/any_iterator.git any_iterator
cd any_iterator/implementation && cmake . && make AnyIteratorTests && ./AnyIteratorTests
This will build the test suite for the sample implementation. Please feel free to add test cases as desired and submit them in a PR, or alternatively, raise any issues that you discover.
Comments on the proposal paper are also welcomed: please raise them as issues on the github page.
An example program using the any_iterator sample implementation might look something like the following:
using namespace sample;
void capitalize(any_input_iterator<std::string> first,
any_input_iterator<std::string> last,
any_output_iterator<std::string> d_outputs)
{
std::transform(first, last, d_outputs, [](auto str) {
// Ignore the `::toupper` potential UB ;)
std::transform(begin(str), end(str), begin(str), ::toupper);
return str;
});
}
int main()
{
capitalize(std::istream_iterator<std::string>(std::cin),
std::istream_iterator<std::string>(),
std::ostream_iterator<std::string>(std::cout, "\n"));
}There are some simple microbenchmarks included within the implementation/benchmarks folder, using Google Benchmark. These are provided to give some perspective on the relative performance of using the iterator directly or using an any_iterator.
These can be built by cloning the repository as above and initializing CMake (preferably using the -DCMAKE_BUILD_TYPE=Release flag) and then calling make AnyIteratorBenchmarks.
On my machine (Intel® Core™ i7-3630QM CPU @ 2.40GHz × 8, 7.7GiB DDR4 RAM) I get the following results:
---------------------------------------------------------------------------------------------------------
Benchmark Time CPU Iterations
---------------------------------------------------------------------------------------------------------
BM_IteratorCreation<sample::any_input_iterator<int>>/200 6 ns 6 ns 111440708
BM_IteratorCreation<ContainerType::iterator>/200 1 ns 1 ns 819729215
BM_IteratorCopy<sample::any_input_iterator<int>>/200 6 ns 6 ns 112859931
BM_IteratorCopy<ContainerType::iterator>/200 0 ns 0 ns 1000000000
BM_IteratorCopyToOutput<sample::any_input_iterator<int>>/200 1267 ns 1266 ns 527775
BM_IteratorCopyToOutput<ContainerType::iterator>/200 23 ns 23 ns 30847233
BM_IteratorOutputIt<sample::any_output_iterator<int>>/200 1356 ns 1355 ns 571798
BM_IteratorOutputIt<std::back_insert_iterator<ContainerType>>/200 761 ns 760 ns 905092
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent