Distribute std::tuple_(size|element) specializations to implementation files

[davebayer]

I want to implement tuple protocol for integer_sequence, but there is a non-trivial circular dependency, because we specialize all relevant std:: traits in one file, which includes make_index_sequence.

So, I decided to decentralize the specializations and move them to the places where we specialize the cuda::std:: traits.

Note that we already specialize for example std::formatter inside the implementation files without any problems.

[coderabbitai]

Note

Reviews paused

It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.

Use the following commands to manage reviews:

• @coderabbitai resume to resume automatic reviews.
• @coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

• ▶️ Resume reviews
• 🔍 Trigger review

📝 Walkthrough

Walkthrough

important: Remove centralized structured_bindings.h; distribute tuple_size/tuple_element specializations into core tuple trait headers and per-type headers, add std ↔ cuda::std interoperability, and update includes and tests.

Changes

Tuple Protocol Refactoring

Layer / File(s)	Summary
Core tuple trait infrastructure libcudacxx/include/cuda/std/__tuple_dir/tuple.h, libcudacxx/include/cuda/std/__tuple_dir/tuple_element.h, libcudacxx/include/cuda/std/__tuple_dir/tuple_size.h, libcudacxx/include/cuda/std/__tuple_dir/tuple_types.h	Core tuple trait headers updated with internal cstddef/type-trait includes, cv-qualified forwarding, std-namespace forwarding specializations, and __tuple_types size/element mappings.
cuda::complex tuple protocol libcudacxx/include/cuda/__complex/complex.h, libcudacxx/include/cuda/std/__complex/complex.h, libcudacxx/include/cuda/std/__complex/tuple.h	Added tuple_size/tuple_element specializations for cuda::complex<_Tp> and host std::complex<_Tp> (guarded), removed duplicate specializations from complex tuple helper header.
cuda::std::pair tuple protocol libcudacxx/include/cuda/std/__utility/pair.h	Reworked includes and added cuda::std tuple_size/tuple_element for ::std::pair (host-guarded) and cuda::std::pair, plus std::tuple_size/std::tuple_element specializations for ::cuda::std::pair.
cuda::std::array tuple protocol libcudacxx/include/cuda/std/array	Replaced prior array trait layout with enable_if(_Ip < _Np)-constrained tuple_element, added host ::std::array support, and added std-namespace specializations for ::cuda::std::array.
cuda::std::tuple tuple protocol libcudacxx/include/cuda/std/__tuple_dir/tuple.h	Added tuple_size/tuple_element specializations for cuda::std::tuple and conditional support for ::std::tuple, and provided std-namespace mappings for cuda::std::tuple.
ranges::subrange tuple protocol libcudacxx/include/cuda/std/__ranges/subrange.h	Rewrote ranges::subrange tuple protocol to use integral_constant, enable_if, and conditional_t for element selection and added std interoperability.
Cleanup and include updates cub/cub/detail/warpspeed/resource/smem_stage.cuh, libcudacxx/include/cuda/std/__tuple_dir/vector_types.h, libcudacxx/include/cuda/std/tuple, thrust/thrust/iterator/detail/tuple_of_iterator_references.h	Removed structured_bindings.h includes, added explicit tuple_element.h/tuple_size.h where required, and adjusted qualification for integral_constant in thrust support.
Tests and test includes libcudacxx/test/libcudacxx/*	Expanded test includes to cover cuda::std::array, cuda::std::complex, and cuda::std::utility; preserved negative-test diagnostics with [[maybe_unused]].

suggested labels: libcu++

suggested reviewers:

• elstehle
• NaderAlAwar

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

♻️ Duplicate comments (1)

libcudacxx/include/cuda/std/__ranges/subrange.h (1)

496-504: ⚠️ Potential issue | 🔴 Critical | ⚡ Quick win

critical: Same inheritance bug as the cuda::std versions—needs type_identity wrapper.

 template <::cuda::std::size_t _Idx, class _Ip, class _Sp, ::cuda::std::ranges::subrange_kind _Kp>
 struct tuple_element<_Idx, ::cuda::std::ranges::subrange<_Ip, _Sp, _Kp>>
-    : ::cuda::std::enable_if_t<(_Idx < 2), ::cuda::std::conditional_t<(_Idx == 0), _Ip, _Sp>>
+    : ::cuda::std::enable_if_t<(_Idx < 2), ::cuda::std::type_identity<::cuda::std::conditional_t<(_Idx == 0), _Ip, _Sp>>>
 {};

 template <::cuda::std::size_t _Idx, class _Ip, class _Sp, ::cuda::std::ranges::subrange_kind _Kp>
 struct tuple_element<_Idx, const ::cuda::std::ranges::subrange<_Ip, _Sp, _Kp>>
-    : ::cuda::std::enable_if_t<(_Idx < 2), ::cuda::std::conditional_t<(_Idx == 0), _Ip, _Sp>>
+    : ::cuda::std::enable_if_t<(_Idx < 2), ::cuda::std::type_identity<::cuda::std::conditional_t<(_Idx == 0), _Ip, _Sp>>>
 {};

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 6eb35ad0-cca1-45fc-a35d-855ab6081716

📥 Commits

Reviewing files that changed from the base of the PR and between e6af9a9 and 2b10101.

📒 Files selected for processing (15)

• cub/cub/detail/warpspeed/resource/smem_stage.cuh
• libcudacxx/include/cuda/__complex/complex.h
• libcudacxx/include/cuda/std/__complex/complex.h
• libcudacxx/include/cuda/std/__complex/tuple.h
• libcudacxx/include/cuda/std/__ranges/subrange.h
• libcudacxx/include/cuda/std/__tuple_dir/structured_bindings.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple_element.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple_size.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple_types.h
• libcudacxx/include/cuda/std/__tuple_dir/vector_types.h
• libcudacxx/include/cuda/std/__utility/pair.h
• libcudacxx/include/cuda/std/array
• libcudacxx/include/cuda/std/tuple
• thrust/thrust/iterator/detail/tuple_of_iterator_references.h

💤 Files with no reviewable changes (4)

• libcudacxx/include/cuda/std/__tuple_dir/structured_bindings.h
• libcudacxx/include/cuda/std/__tuple_dir/vector_types.h
• libcudacxx/include/cuda/std/tuple
• libcudacxx/include/cuda/std/__complex/tuple.h

[coderabbitai]

Actionable comments posted: 3

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: af568c0e-a7ad-4440-aa0e-87981093659f

📥 Commits

Reviewing files that changed from the base of the PR and between e29e5e9 and 41f3cca.

📒 Files selected for processing (18)

• cub/cub/detail/warpspeed/resource/smem_stage.cuh
• libcudacxx/include/cuda/__complex/complex.h
• libcudacxx/include/cuda/std/__complex/complex.h
• libcudacxx/include/cuda/std/__complex/tuple.h
• libcudacxx/include/cuda/std/__ranges/subrange.h
• libcudacxx/include/cuda/std/__tuple_dir/structured_bindings.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple_element.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple_size.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple_types.h
• libcudacxx/include/cuda/std/__tuple_dir/vector_types.h
• libcudacxx/include/cuda/std/__utility/pair.h
• libcudacxx/include/cuda/std/array
• libcudacxx/include/cuda/std/tuple
• libcudacxx/test/libcudacxx/cuda/utilities/tuple/std_types_tuple_element.pass.cpp
• libcudacxx/test/libcudacxx/cuda/utilities/tuple/std_types_tuple_size.pass.cpp
• libcudacxx/test/libcudacxx/std/containers/sequences/array/array.tuple/tuple_element.fail.cpp
• thrust/thrust/iterator/detail/tuple_of_iterator_references.h

💤 Files with no reviewable changes (4)

• libcudacxx/include/cuda/std/__tuple_dir/vector_types.h
• libcudacxx/include/cuda/std/tuple
• libcudacxx/include/cuda/std/__tuple_dir/structured_bindings.h
• libcudacxx/include/cuda/std/__complex/tuple.h

✅ Files skipped from review due to trivial changes (1)

• libcudacxx/test/libcudacxx/std/containers/sequences/array/array.tuple/tuple_element.fail.cpp

[coderabbitai]

Actionable comments posted: 1

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 5c25d0eb-118b-4488-be2d-4a6744c8b6b7

📥 Commits

Reviewing files that changed from the base of the PR and between 41f3cca and 91c17e9.

📒 Files selected for processing (18)

• cub/cub/detail/warpspeed/resource/smem_stage.cuh
• libcudacxx/include/cuda/__complex/complex.h
• libcudacxx/include/cuda/std/__complex/complex.h
• libcudacxx/include/cuda/std/__complex/tuple.h
• libcudacxx/include/cuda/std/__ranges/subrange.h
• libcudacxx/include/cuda/std/__tuple_dir/structured_bindings.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple_element.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple_size.h
• libcudacxx/include/cuda/std/__tuple_dir/tuple_types.h
• libcudacxx/include/cuda/std/__tuple_dir/vector_types.h
• libcudacxx/include/cuda/std/__utility/pair.h
• libcudacxx/include/cuda/std/array
• libcudacxx/include/cuda/std/tuple
• libcudacxx/test/libcudacxx/cuda/utilities/tuple/std_types_tuple_element.pass.cpp
• libcudacxx/test/libcudacxx/cuda/utilities/tuple/std_types_tuple_size.pass.cpp
• libcudacxx/test/libcudacxx/std/containers/sequences/array/array.tuple/tuple_element.fail.cpp
• thrust/thrust/iterator/detail/tuple_of_iterator_references.h

💤 Files with no reviewable changes (4)

• libcudacxx/include/cuda/std/tuple
• libcudacxx/include/cuda/std/__tuple_dir/structured_bindings.h
• libcudacxx/include/cuda/std/__tuple_dir/vector_types.h
• libcudacxx/include/cuda/std/__complex/tuple.h

✅ Files skipped from review due to trivial changes (1)

• cub/cub/detail/warpspeed/resource/smem_stage.cuh

[github-actions]

🥳 CI Workflow Results

🟩 Finished in 3h 10m: Pass: 100%/398 | Total: 19d 22h | Max: 3h 09m | Hits: 24%/4225373

See results here.

[miscco]

Looks good, I believe with the added forward declarations this is now the better approach