GH-45362: [C++] Fix identity cast for time and list scalar

[mapleFU]

Rationale for this change

The PR #40237 introduced code:

// time to time
template <typename To, typename From, typename T = typename To::TypeClass>

However, the Time64Type::TypeClass doesn't exist, so SFINAE always failed.

Are these changes tested?

Yes

Are there any user-facing changes?

No.

• GitHub Issue: [C++]: Identity cast for Time64Scalar and Time32Scalar fails #45362

[github-actions]

⚠️ GitHub issue #45362has been automatically assigned in GitHub to PR creator.

[mapleFU]

Windows CI failed seems to be related.

[mapleFU]

Also cc @pitrou since this is new year and @zanmato1984 might not have time for this

[mapleFU]

decimal seems not have castTo

[pitrou]

Can we open issues for the missing casts?

[mapleFU]

Added, but seems previous issue wants to deprecate them, I think these issues may also be mark as not planned, lol

[mapleFU]

struct and map seems not supports castTo non string types

[pitrou]

Thanks for the update @mapleFU !

[pitrou]

Can we open issues for the missing casts?

[pitrou]

Thanks for the update @mapleFU ! LGTM.

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 4 benchmarking runs that have been run so far on merge-commit d9e7fa6.

There were 8 benchmark results with an error:

• Commit Run on arm64-t4g-2xlarge-linux at 2025-02-05 15:07:01Z
  • tpch (R) with engine=arrow, format=native, language=R, memory_map=False, query_id=TPCH-03, scale_factor=1
  • tpch (R) with engine=arrow, format=parquet, language=R, memory_map=False, query_id=TPCH-03, scale_factor=1
• and 6 more (see the report linked below)

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 4 possible false positives for unstable benchmarks that are known to sometimes produce them.