[davebayer]

This PR implements ranges::__static_size_v traits that holds the size of the range if it can be statically determined, dynamic_extent otherwise.

[coderabbitai]

Important

Review skipped

Review was skipped as selected files did not have any reviewable changes.

Run configuration

• Source mode: Auto
• Repository: NVIDIA/cccl
• Branch: main
• File: .coderabbit.yaml
• Commit: 5b5249e
• Layers: Repository YAML
• Review profile: CHILL
• Plan: Enterprise
• Run ID: 48991e00-3486-4854-bc6a-d3d968efdfab

Commits

Reviewing files that changed from the base of the PR and between e525bcf and 239dde8.

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• Trigger review

No actionable comments were generated in the recent review.

Summary

Summary by CodeRabbit

• New Features
  • Added compile-time size information for ranges with statically known sizes, including built-in arrays, cuda::std::array, fixed-extent spans, and supported custom ranges. Ranges with dynamic or unavailable sizes are identified as having a dynamic extent.
• Tests
  • Added compile-time coverage for supported range types, cv/ref variations, and dynamic or unavailable sizes, including ranges with non-default-constructible element types.

The change adds cuda::std::ranges::__static_size_v. It yields a compile-time size when ranges::size is a size_t constant expression, and otherwise yields dynamic_extent. Compile-time tests cover built-in and custom range types.

Changes: Static Range Size

Layer / File(s)	Summary
Static size computation libcudacxx/include/cuda/std/__ranges/size.h	The header adds dependencies and defines __static_size_v to compute a static size or return dynamic_extent.
Compile-time validation libcudacxx/test/libcudacxx/std/ranges/range.access/static_size.compile.pass.cpp	The tests check fixed-size and dynamic-size ranges, unsupported cases, cv/ref forms, and the variable’s result type.

Change: Feature

Merge Risk: Minimal · previous review e525b

This adds a new internal trait, ranges::__static_size_v, and compile-time tests for it. No existing behavior changes, and no merge-blocking risk was found.

Architecture risk: Low · up to 0bfc3

The change affects 1 system.

Changed systems: libcudacxx

Architecture concerns
No architecture-level concerns identified.

Review details

Systems and components

• observed — libcudacxx (service) was modified; 2 changed files map to changed impact.

Before / after behavior

• observed — Modified behavior in libcudacxx/test/libcudacxx/std/ranges/range.access/static_size.compile.pass.cpp: Adds headers and test types representing fixed-size and dynamic-size ranges, including member-size, ADL-size, runtime-size, value-dependent-size, sentinel-sized, and disabled-sized cases.
• observed — Modified behavior in libcudacxx/test/libcudacxx/std/ranges/range.access/static_size.compile.pass.cpp: Adds test_cvref, which asserts the expected static size for mutable, const, volatile, and reference-qualified forms of a type, and verifies the result type is const cuda::std::size_t.
• observed — Modified behavior in libcudacxx/test/libcudacxx/std/ranges/range.access/static_size.compile.pass.cpp: Adds a compile-time test covering expected sizes for arrays, cuda::std::array, fixed-extent spans, and custom fixed-size ranges; it checks dynamic_extent for dynamic or unsupported cases and runs the test from main.
• observed — Modified behavior in libcudacxx/include/cuda/std/__ranges/size.h: The header adds the span forward declaration and includes integral_constant and void_t; the other listed includes remain in place.

- [ ] Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Recent review info

Run configuration

• Source mode: Auto
• Repository: NVIDIA/cccl
• Branch: main
• File: .coderabbit.yaml
• Commit: 5060e9b
• Layers: Repository YAML
• Review profile: CHILL
• Plan: Enterprise
• Run ID: 03e0f518-c6b0-4bf3-b405-265515bf2d37

Commits

Reviewing files that changed from the base of the PR and between 0bfc31e and e525bcf.

Files selected for processing: 1

• libcudacxx/include/cuda/std/__ranges/size.h

Files skipped from review as they are similar to previous changes: 1

• libcudacxx/include/cuda/std/__ranges/size.h

Included review availability: This review used your included allowance. Your plan provides up to 12 included reviews per hour; 11 remain after this review.

[Jacobfaib]

declval?

[davebayer]

It can't be declval, because it would be invoked, triggering the static_assertion inside it

[Jacobfaib]

Why extract the ::type here? void_t<integral_constant> is enough to exercise the SFINAE. Or I guess void_t<decltype(range::size(...))> would do it too.

[davebayer]

void_t<decltype(range::size(...))> wouldn't be enough, we need the value of the invocation, not the return type. We query ::type because we need to make sure that the range::size(...) is a constant expression and instantiates integral_constant

[github-actions]

🥳 CI Workflow Results

🟩 Finished in 5h 50m: Pass: 100%/230 | Total: 9d 18h | Max: 5h 50m | Hits: 30%/4260751

See results here.