perf: optimize spark_regexp_extract_all in datafusion-comet-spark-expr

[andygrove]

This PR was created by an LLM as a draft PR. I will mark it as ready for review after human review.

Which issue does this PR close?

N/A — autonomous exploratory PR.

Rationale for this change

Reuse a single CaptureLocations across all matches and rows instead of letting captures_iter heap-allocate capture slots per match, and use find_iter when only the whole match (idx=0) is needed.

What changes are included in this PR?

Reuse a single CaptureLocations across all matches and rows instead of letting captures_iter heap-allocate capture slots per match, and use find_iter when only the whole match (idx=0) is needed.

How are these changes tested?

Correctness: unit tests + seeded differential fuzz (bit-identical Arrow output vs main).

Benchmark (criterion):

• capture_group: 24.244% faster (base 1052449ns -> cand 797294ns)
• whole_match: 61.571% faster (base 1046798ns -> cand 402278ns)

Full criterion output:

regexp_extract_all/capture_group
                        time:   [797.09 µs 797.74 µs 798.40 µs]
                        change: [−24.419% −24.244% −24.094%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild
regexp_extract_all/whole_match
                        time:   [402.08 µs 402.21 µs 402.42 µs]
                        change: [−61.621% −61.571% −61.519%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 8 outliers among 100 measurements (8.00%)
  1 (1.00%) low mild
  3 (3.00%) high mild
  4 (4.00%) high severe