perf: optimize to_csv_inner 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

Replaced up to six per-field full-string pattern searches and char-by-char escaping with one precomputed ASCII byte scan and bulk run copies, keeping the original searches as a fallback for non-ASCII/multi-char delimiters.

What changes are included in this PR?

Replaced up to six per-field full-string pattern searches and char-by-char escaping with one precomputed ASCII byte scan and bulk run copies, keeping the original searches as a fallback for non-ASCII/multi-char delimiters.

How are these changes tested?

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

Benchmark (criterion):

• to_csv: 7.281% faster (base 858808ns -> cand 796276ns)

Full criterion output:

to_csv                  time:   [791.73 µs 793.59 µs 795.87 µs]
                        change: [−7.5817% −7.2812% −6.9754%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 6 outliers among 100 measurements (6.00%)
  1 (1.00%) low mild
  4 (4.00%) high mild
  1 (1.00%) high severe