[libc++] Implement std::inplace_vector<T, N>
#105981
Draft
Conversation
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
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
|
To-do list:
|
MitalAshok
force-pushed
the
inplace-vector
branch
from
a4127b3
to
22903da
Compare
MitalAshok
force-pushed
the
inplace-vector
branch
from
22903da
to
47615ff
Compare
MitalAshok
force-pushed
the
inplace-vector
branch
8 times, most recently
from
91e3d83
to
8d29e36
Compare
MitalAshok
force-pushed
the
inplace-vector
branch
from
8d29e36
to
440c327
Compare
jiixyj
reviewed
Oct 26, 2024
| } | ||
| template <_ContainerCompatibleRange<_Tp> _Range> | ||
| constexpr _LIBCPP_HIDE_FROM_ABI inplace_vector(from_range_t, _Range&& __rg) : inplace_vector() { | ||
| append_range<true>(__rg); |
There was a problem hiding this comment.
Suggested change
| append_range<true>(__rg); | |
| append_range(std::forward<_Range>(__rg)); |
This line does not compile on my machine. I guess this was supposed to forward to the append_range member function?
philnik777
reviewed
Nov 8, 2024
Comment on lines
+227
to
+231
| if constexpr (_MaximumPadding >= sizeof(size_t)) { | ||
| // Unconditionally use size_t in this case. Using a smaller size | ||
| // will not change the size of the inplace_vector because it will | ||
| // only introduce more padding bits | ||
| return static_cast<size_t>(0); |
There was a problem hiding this comment.
Not using size_t would allow others (e.g. expected) to use the padding bytes though, so we should still use a smaller type.
| // will not change the size of the inplace_vector because it will | ||
| // only introduce more padding bits | ||
| return static_cast<size_t>(0); | ||
| } else if constexpr (_Capacity <= static_cast<unsigned char>(-1) && (sizeof(unsigned char) < sizeof(size_t))) { |
There was a problem hiding this comment.
Suggested change
| } else if constexpr (_Capacity <= static_cast<unsigned char>(-1) && (sizeof(unsigned char) < sizeof(size_t))) { | |
| } else if constexpr (_Capacity <= numeric_limits<unsigned char>::max() && (sizeof(unsigned char) < sizeof(size_t))) { |
Why do you check for sizeof(T) < sizeof(size_t)? It's not like we care in the end which type exactly is used.
Comment on lines
+218
to
+220
| _LIBCPP_BEGIN_NAMESPACE_STD | ||
|
|
||
| #if _LIBCPP_STD_VER >= 26 |
There was a problem hiding this comment.
Suggested change
| _LIBCPP_BEGIN_NAMESPACE_STD | |
| #if _LIBCPP_STD_VER >= 26 | |
| #if _LIBCPP_STD_VER >= 26 | |
| _LIBCPP_BEGIN_NAMESPACE_STD |
| } | ||
|
|
||
| template <class _Tp, size_t _Capacity> | ||
| class _LIBCPP_TEMPLATE_VIS __inplace_vector_array { |
There was a problem hiding this comment.
Suggested change
| class _LIBCPP_TEMPLATE_VIS __inplace_vector_array { | |
| class __inplace_vector_array { |
It's and internal type, so it shouldn't need any annotations.
| } | ||
|
|
||
| template <class _Tp, size_t _Capacity> | ||
| class _LIBCPP_TEMPLATE_VIS __inplace_vector_array { |
There was a problem hiding this comment.
Why did you introduce a base class?
Comment on lines
+596
to
+612
| constexpr _LIBCPP_HIDE_FROM_ABI iterator begin() noexcept { return __make_iter(data()); } | ||
| constexpr _LIBCPP_HIDE_FROM_ABI const_iterator begin() const noexcept { return __make_iter(data()); } | ||
| constexpr _LIBCPP_HIDE_FROM_ABI iterator end() noexcept { return __make_iter(data() + size()); } | ||
| constexpr _LIBCPP_HIDE_FROM_ABI const_iterator end() const noexcept { return __make_iter(data() + size()); } | ||
| constexpr _LIBCPP_HIDE_FROM_ABI reverse_iterator rbegin() noexcept { return reverse_iterator(end()); } | ||
| constexpr _LIBCPP_HIDE_FROM_ABI const_reverse_iterator rbegin() const noexcept { | ||
| return const_reverse_iterator(end()); | ||
| } | ||
| constexpr _LIBCPP_HIDE_FROM_ABI reverse_iterator rend() noexcept { return reverse_iterator(begin()); } | ||
| constexpr _LIBCPP_HIDE_FROM_ABI const_reverse_iterator rend() const noexcept { | ||
| return const_reverse_iterator(begin()); | ||
| } | ||
|
|
||
| constexpr _LIBCPP_HIDE_FROM_ABI const_iterator cbegin() const noexcept { return begin(); } | ||
| constexpr _LIBCPP_HIDE_FROM_ABI const_iterator cend() const noexcept { return end(); } | ||
| constexpr _LIBCPP_HIDE_FROM_ABI const_reverse_iterator crbegin() const noexcept { return rbegin(); } | ||
| constexpr _LIBCPP_HIDE_FROM_ABI const_reverse_iterator crend() const noexcept { return rend(); } |
There was a problem hiding this comment.
These should probably all be [[nodiscard]].
|
|
||
| template <class _First, class _Sent> | ||
| constexpr _LIBCPP_HIDE_FROM_ABI void __assign_with_size(_First&& __first, _Sent&& __sent, size_t __n) { | ||
| if (__n > capacity()) [[unlikely]] |
There was a problem hiding this comment.
Suggested change
| if (__n > capacity()) [[unlikely]] | |
| if (__n > capacity()) |
You call a [[noreturn]] function in this branch. The compiler already marks that us unlikely. No need to do it explicitly.
| union { | ||
| _Tp __elems_[_Capacity]; | ||
| }; | ||
| __small_size_type __sz_; |
There was a problem hiding this comment.
I think we should look into using #98498 here. It could be a really sweet optimization, but I don't know how much more code would be generated.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
P0843R14 https://wg21.link/P0843R14:
std::inplace_vector<T, N>The ABI of this class is an array of
T[N]followed by a__small_size_typemember to hold the size, which is some unsigned integral type depending onTandN(at least large enough to hold up toN). Or ifN == 0, an empty class as mandated by the standard.This is not the fastest implementation. It aims to be correct and able to be optimised in the future in an ABI compatible way.
Closes #105433