Make is_supported_cast fail for variable-width inputs

[JohnZed]

Description

cudf is_supported_cast returns true for string inputs. But cudf cast documentation says that it does not support variable-width
types.

This leads to an issue in velox (facebookincubator/velox#18450) where the engine thinks it can
use cudf cast to convert varchars to ints, though that will ultimately need an exception. Velox will eventually need to use
a different strings API for this conversion, but at least changing is_supported_cast will allow it to no longer route conversions
through this path.

Checklist

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

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[bdice]

Thanks!

[PointKernel]

pre-commit.ci autofix

[PointKernel]

/ok to test 52419b5

[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: 539668b2-eab0-418d-bbb0-1c5bf34fd1d3

📥 Commits

Reviewing files that changed from the base of the PR and between 52419b5 and 34b1762.

📒 Files selected for processing (1)

• cpp/tests/unary/cast_tests.cpp

🚧 Files skipped from review as they are similar to previous changes (1)

• cpp/tests/unary/cast_tests.cpp

---

📝 Walkthrough

Summary by CodeRabbit

• Bug Fixes

  • Corrected cast support validation to consistently require fixed-width source and target types.
  • Casting from STRING to INT32 is now correctly reported as unsupported.

• Tests

  • Added coverage verifying unsupported string-to-integer cast behavior.

Walkthrough

The cast support check now requires fixed-width source and target types. A unit test verifies that casting from STRING to INT32 is unsupported. SPDX notices were updated.

Changes

Cast support validation

Layer / File(s)	Summary
Cast predicate and regression test cpp/src/unary/cast_ops.cu, cpp/tests/unary/cast_tests.cpp	is_supported_non_fixed_point_cast now checks both source and target fixed-width types. A test verifies that STRING to INT32 is unsupported. SPDX notices were updated.

Estimated code review effort: 2 (Simple) | ~10 minutes

Possibly related PRs

• rapidsai/cudf#23462: Both PRs modify cast-type validation and unsupported cast behavior in different APIs and code paths.

Suggested reviewers: abigalekim, igorpeshansky

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly summarizes the main change: rejecting variable-width inputs in is_supported_cast.
Description check	✅ Passed	The description explains the variable-width input issue, its Velox impact, and the included test coverage.
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.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[coderabbitai]

🧹 Nitpick comments (1)

cpp/tests/unary/cast_tests.cpp (1)

204-208: 🎯 Functional Correctness | 🔵 Trivial | ⚡ Quick win

Cover variable-width targets too.

The implementation checks both From and To, but this test only checks a variable-width source. Add an INT32 to STRING assertion so a regression in the target check cannot pass unnoticed.

Proposed test extension

+  EXPECT_FALSE(cudf::is_supported_cast(cudf::data_type{cudf::type_id::INT32},
+                                       cudf::data_type{cudf::type_id::STRING}));

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@cpp/tests/unary/cast_tests.cpp` around lines 204 - 208, Extend the
StringToInt32IsUnsupported test to also assert that an INT32 source to STRING
target is unsupported, covering the variable-width target path while preserving
the existing assertion.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@cpp/tests/unary/cast_tests.cpp`:
- Around line 204-208: Extend the StringToInt32IsUnsupported test to also assert
that an INT32 source to STRING target is unsupported, covering the
variable-width target path while preserving the existing assertion.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: d85c5012-f405-490c-8203-2bd7700e60d6

📥 Commits

Reviewing files that changed from the base of the PR and between 463fba2 and 52419b5.

📒 Files selected for processing (2)

• cpp/src/unary/cast_ops.cu
• cpp/tests/unary/cast_tests.cpp

[JohnZed]

/ok to test 34b1762

[JohnZed]

/ok to test 2b79f91