Change deprecated calls of stream_ref.wait() to stream_ref.sync()

[davidwendt]

Description

Fixes deprecation warnings for calls to cuda::stream_ref::wait() by changing them to the recommended cuda::stream_ref::sync() instead.

Example deprecation warning:

/cudf/cpp/src/partitioning/partitioning.cu:745:12: warning: 'void cuda::__4::stream_ref::wait() const' is deprecated: Use sync() instead. [-Wdeprecated-declarations]
  745 |     stream.wait();  // Async D2H copy must finish before returning host vec
      |     ~~~~~~~^~
/cudf/cpp/build/_deps/cccl-src/lib/cmake/libcudacxx/../../../libcudacxx/include/cuda/__stream/stream_ref.h:179:95: note: declared here
  179 |   CCCL_DEPRECATED_BECAUSE("Use sync() instead.") _CCCL_HOST_API void wait() const

Introduced by #23645

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 1507cb2b-60e2-4304-8775-26309f56798b

📥 Commits

Reviewing files that changed from the base of the PR and between 0b57a3b and 0e64479.

📒 Files selected for processing (3)

• cpp/src/copying/contiguous_split.cu
• cpp/src/partitioning/partitioning.cu
• cpp/tests/copying/concatenate_tests.cpp

---

📝 Walkthrough

Summary by CodeRabbit

• Bug Fixes
  • Improved synchronization during data copying and hash partitioning operations.
  • Ensured host-side partition metadata is fully available before results are returned.
  • Improved reliability of asynchronous column data and mask transfers.
• Tests
  • Updated copy-related validation to reflect the improved synchronization behavior.

Walkthrough

Changes

CUDA stream synchronization

Layer / File(s)	Summary
Replace stream wait calls with sync calls cpp/src/copying/contiguous_split.cu, cpp/src/partitioning/partitioning.cu, cpp/tests/copying/concatenate_tests.cpp	Production paths and the test fixture now call stream.sync() instead of stream.wait() before accessing copied metadata, partition offsets, or initialized device data.

Estimated code review effort: 1 (Trivial) | ~3 minutes

Mergeability Score: ⚪ Minimal · up to 0e644

This is a localized deprecated API replacement with no actionable merge-blocking risk remaining after normal checks and review.

Suggested reviewers: pmattione-nvidia, qbacpey, vyasr

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Title check	✅ Passed	The title clearly and concisely describes replacing deprecated stream_ref.wait() calls with stream_ref.sync().
Description check	✅ Passed	The description explains the deprecation warning, the replacement, and the affected API calls.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[bdice]

Sorry. I did think about this deprecation when I reviewed #23645, but I forgot which method was deprecated. cc: @vyasr

[vyasr]

We need to audit our CI. I did think about this but I thought I got all the cases, and I trusted CI passing as sufficient because we build with warnings as errors. I don't yet understand how my PR could have passed CI in the current state unless CCCL has an extra knob for its own deprecation warnings somehow.

[davidwendt]

/merge