[array] add c++20 initialization_list constructor

[karczex]

Since C++20 classes with explicitly defaulted constructors are no more
aggregates, so constructor for initialization_list is needed.

---

This change is 

[codecov]

Codecov Report

Merging #1158 (afa93e0) into master (0ddba32) will increase coverage by 0.01%.
The diff coverage is 50.00%.

@@            Coverage Diff             @@
##           master    #1158      +/-   ##
==========================================
+ Coverage   94.05%   94.06%   +0.01%     
==========================================
  Files          52       52              
  Lines        5163     5124      -39     
==========================================
- Hits         4856     4820      -36     
+ Misses        307      304       -3     

Flag	Coverage Δ
tests_clang_debug_cpp17	93.63% <50.00%> (+0.08%)	⬆️
tests_gcc_debug	90.66% <100.00%> (-0.14%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
include/libpmemobj++/container/array.hpp	98.07% <50.00%> (+0.42%)	⬆️
include/libpmemobj++/experimental/radix_tree.hpp	98.46% <0.00%> (+0.14%)	⬆️
include/libpmemobj++/detail/life.hpp	97.43% <0.00%> (+0.21%)	⬆️
include/libpmemobj++/transaction.hpp	82.35% <0.00%> (+0.84%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0ddba32...afa93e0. Read the comment docs.

[igchor]

In future, we might think about implementing something like was proposed here: https://stackoverflow.com/questions/38932089/can-i-initialize-an-array-using-the-stdinitializer-list-instead-of-brace-enclo to avoid the limitations of std::initializer_list

Reviewable status: 0 of 6 files reviewed, 2 unresolved discussions (waiting on @karczex)

---

include/libpmemobj++/container/array.hpp, line 49 at r1 (raw file):

 * aggregate since C++20.
 *
 * @note Since C++20 pmem::obj::array cannot be used for types with deleted

I don't think this is true - you can use array in that case, you just cannot initialize those values in a ctor. You can just do (if T has a default ctor):

pmem::obj::array<T, 3> arr;
arr[0] = ...;

---

include/libpmemobj++/container/array.hpp, line 52 at r1 (raw file):

 * copy constructor. This may change in the future.
 *
 * @note Since C++20 pmem::obj::array cannot be used for const types. This may

Well, it can, but you can't use any ctor different than a default one.

[lukaszstolarczuk]

Reviewable status: 0 of 6 files reviewed, 6 unresolved discussions (waiting on @karczex)

---

include/libpmemobj++/container/array.hpp, line 46 at r1 (raw file):

 * to a transaction while traversing.
 *
 * @note Accordingly to changes in C++ standard, pmem::obj::array is not an

According to

---

tests/CMakeLists.txt, line 357 at r1 (raw file):

		add_test_generic(NAME array_slice_pmreorder CASE 1 TRACERS none SCRIPT array/array_slice_pmreorder_1.cmake)
	endif()
	if(CXX_STANDARD GREATER_EQUAL 17)

GREATER_EQUAL is delivered in CMake >= 3.7, so please just use CXX_STANDARD GREATER 14

---

tests/array/array_is_aggregate.cpp, line 8 at r1 (raw file):

/* Test which checks if array behave in the same way as any structure with
 * defaulted constructor - accordingly to used c++ standard */

according to

---

tests/external/CMakeLists.txt, line 108 at r1 (raw file):

	if(CXX_STANDARD LESS_EQUAL 17)

. (LESS 20)

[KFilipek]

Reviewed 6 of 6 files at r1.
Reviewable status: all files reviewed, 13 unresolved discussions (waiting on @karczex and @lukaszstolarczuk)

---

include/libpmemobj++/container/array.hpp, line 15 at r1 (raw file):

#include <functional>

#include <initializer_list>

one block up:

#include <algorithm>
#include <functional>
#include <initializer_list>

---

tests/CMakeLists.txt, line 357 at r1 (raw file):

Previously, lukaszstolarczuk (Łukasz Stolarczuk) wrote…

GREATER_EQUAL is delivered in CMake >= 3.7, so please just use CXX_STANDARD GREATER 14

or update README.md, on every supported OS this PR was built successfully

---

tests/array/array_is_aggregate.cpp, line 4 at r1 (raw file):

#include "unittest.hpp"
#include <libpmemobj++/container/array.hpp>

suparate this into 2 blocks

---

tests/array/array_is_aggregate.cpp, line 21 at r1 (raw file):

		std::is_aggregate<struct_with_defaulted_constructor>::value;

	UT_ASSERTeq(array_is_aggregate, cpp_standard_support);

@igchor is static_assert here better than UT_ASSERTeq?

---

tests/external/CMakeLists.txt, line 107 at r1 (raw file):

	add_test_generic(NAME array_get_const TRACERS none)

2 lines instead of 1

---

tests/external/libcxx/array/array.cons/implicit_copy_const.pass.cpp, line 10 at r1 (raw file):

//===----------------------------------------------------------------------===//
//
// Copyright 2018-2021, Intel Corporation

If it's new file it should contain only 2021

---

tests/external/libcxx/array/array.cons/implicit_copy_const.pass.cpp, line 54 at r1 (raw file):

	run()
	{
		((void)c2);

(void)c2; should be sufficient

---

tests/external/libcxx/array/array.cons/implicit_copy_const.pass.cpp, line 67 at r1 (raw file):

	run()
	{
		((void)c2);

.

[lukaszstolarczuk]

Reviewable status: all files reviewed, 13 unresolved discussions (waiting on @karczex and @KFilipek)

---

tests/CMakeLists.txt, line 357 at r1 (raw file):

Previously, KFilipek (Krzysztof Filipek) wrote…

or update README.md, on every supported OS this PR was built successfully

it may break on OSes like RHEL - and what would be the gain...? I'd say, it's better to use as low version of CMake as possible 😉

---

tests/external/libcxx/array/array.cons/implicit_copy_const.pass.cpp, line 10 at r1 (raw file):

Previously, KFilipek (Krzysztof Filipek) wrote…

If it's new file it should contain only 2021

nope, as usual - we copy-pasted the content from previously existing file - hence both dates

[lukaszstolarczuk]

👍

Reviewed 2 of 6 files at r1, 8 of 8 files at r2.
Reviewable status: all files reviewed, 10 unresolved discussions (waiting on @karczex and @KFilipek)

---

utils/docker/run-build.sh, line 217 at r2 (raw file):

	printf "\n$(tput setaf 1)$(tput setab 7)BUILD ${FUNCNAME[0]} START$(tput sgr 0)\n"
	mkdir build && cd build

we may add here the check for support in CMake (like here: https://github.com/pmem/pmemkv/blob/master/utils/docker/run-build.sh#L168), but it will work on most OSes...

[KFilipek]

Reviewable status: all files reviewed, 9 unresolved discussions (waiting on @karczex)

---

tests/CMakeLists.txt, line 357 at r1 (raw file):

Previously, lukaszstolarczuk (Łukasz Stolarczuk) wrote…

it may break on OSes like RHEL - and what would be the gain...? I'd say, it's better to use as low version of CMake as possible 😉

Indeed

[igchor]

Reviewable status: all files reviewed, 6 unresolved discussions (waiting on @karczex and @KFilipek)

---

tests/array/array_is_aggregate.cpp, line 21 at r1 (raw file):

Previously, KFilipek (Krzysztof Filipek) wrote…

@igchor is static_assert here better than UT_ASSERTeq?

I think it's fine both ways.

---

tests/external/CMakeLists.txt, line 107 at r1 (raw file):

Previously, KFilipek (Krzysztof Filipek) wrote…

2 lines instead of 1

Done.

---

tests/external/libcxx/array/array.cons/implicit_copy_const.pass.cpp, line 54 at r1 (raw file):

Previously, KFilipek (Krzysztof Filipek) wrote…

(void)c2; should be sufficient

Done.

---

tests/external/libcxx/array/array.cons/implicit_copy_const.pass.cpp, line 67 at r1 (raw file):

Previously, KFilipek (Krzysztof Filipek) wrote…

.

.

---

utils/docker/run-build.sh, line 217 at r2 (raw file):

Previously, lukaszstolarczuk (Łukasz Stolarczuk) wrote…

we may add here the check for support in CMake (like here: https://github.com/pmem/pmemkv/blob/master/utils/docker/run-build.sh#L168), but it will work on most OSes...

Done.

[lukaszstolarczuk]

Reviewed 4 of 4 files at r3.
Reviewable status: all files reviewed, 9 unresolved discussions (waiting on @karczex and @KFilipek)

---

include/libpmemobj++/container/array.hpp, line 116 at r3 (raw file):

	 *
	 * It provides similar initialization semantics for
	 * C++20 (where pmem::obj::array is not longer an aggregate).

no longer

---

tests/external/libcxx/array/array.cons/implicit_copy_const.pass.cpp, line 67 at r1 (raw file):

Previously, igchor (Igor Chorążewicz) wrote…

.

nope..?

---

tests/external/libcxx/array/array.cons/initializer_list.pass.cpp, line 56 at r3 (raw file):

	C c = {1};
#else
	/* This syntax casuse warning: braces around scalar initializer

casuse -> causes

---

utils/docker/run-build.sh, line 217 at r3 (raw file):

	printf "\n$(tput setaf 1)$(tput setab 7)BUILD ${FUNCNAME[0]} START$(tput sgr 0)\n"

	if [ ${CMAKE_VERSION_NUMBER} -lt 312 ]; then

but we never set CMAKE_VERSION_NUMBER 😉 (in pmemkv it's done in prepare-for-build.sh)

[igchor]

Reviewable status: 5 of 11 files reviewed, 9 unresolved discussions (waiting on @KFilipek and @lukaszstolarczuk)

---

include/libpmemobj++/container/array.hpp, line 15 at r1 (raw file):

Previously, KFilipek (Krzysztof Filipek) wrote…

one block up:

#include <algorithm>
#include <functional>
#include <initializer_list>

Done.

---

include/libpmemobj++/container/array.hpp, line 116 at r3 (raw file):

Previously, lukaszstolarczuk (Łukasz Stolarczuk) wrote…

no longer

Done.

---

tests/array/array_is_aggregate.cpp, line 4 at r1 (raw file):

Previously, KFilipek (Krzysztof Filipek) wrote…

#include "unittest.hpp"
#include <libpmemobj++/container/array.hpp>

suparate this into 2 blocks

Done.

---

tests/external/libcxx/array/array.cons/implicit_copy_const.pass.cpp, line 67 at r1 (raw file):

Previously, lukaszstolarczuk (Łukasz Stolarczuk) wrote…

nope..?

Done.

---

tests/external/libcxx/array/array.cons/initializer_list.pass.cpp, line 56 at r3 (raw file):

Previously, lukaszstolarczuk (Łukasz Stolarczuk) wrote…

casuse -> causes

Done.

---

utils/docker/run-build.sh, line 217 at r3 (raw file):

Previously, lukaszstolarczuk (Łukasz Stolarczuk) wrote…

but we never set CMAKE_VERSION_NUMBER 😉 (in pmemkv it's done in prepare-for-build.sh)

Done.

[lukaszstolarczuk]

Reviewed 4 of 5 files at r4, 2 of 2 files at r5.
Reviewable status: all files reviewed, 6 unresolved discussions (waiting on @KFilipek)