unexpand: improve performances (1.50 faster than GNU)

[sylvestre]

hyperfine --warmup 3 "/usr/bin/unexpand moby64.txt"                     "./target/release/coreutils.prev unexpand moby64.txt" "./target/release/coreutils unexpand moby64.txt"
Benchmark 1: /usr/bin/unexpand moby64.txt
  Time (mean ± σ):      95.8 ms ±   3.8 ms    [User: 81.1 ms, System: 14.6 ms]
  Range (min … max):    89.6 ms … 104.8 ms    31 runs

Benchmark 2: ./target/release/coreutils.prev unexpand moby64.txt
  Time (mean ± σ):     952.3 ms ±   3.0 ms    [User: 846.4 ms, System: 105.6 ms]
  Range (min … max):   949.4 ms … 958.1 ms    10 runs

Benchmark 3: ./target/release/coreutils unexpand moby64.txt
  Time (mean ± σ):      63.7 ms ±   2.2 ms    [User: 52.9 ms, System: 10.7 ms]
  Range (min … max):    60.8 ms …  70.2 ms    46 runs

Summary
  ./target/release/coreutils unexpand moby64.txt ran
    1.50 ± 0.08 times faster than /usr/bin/unexpand moby64.txt
   14.95 ± 0.53 times faster than ./target/release/coreutils.prev unexpand moby64.txt

[sylvestre]

we are still slower with -a

[codspeed-hq]

CodSpeed Performance Report

Merging #8815 will improve performances by 5.35%

_{Comparing sylvestre:perf-unexpand (687716c) with main (02312bf)}

Summary

⚡ 2 improvements
✅ 77 untouched
⏩ 73 skipped¹

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
⚡	unexpand_large_file[10]	577.3 ms	548 ms	+5.35%
⚡	unexpand_many_lines[100000]	275.4 ms	261.4 ms	+5.35%

Footnotes

1. 73 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[github-actions]

GNU testsuite comparison:

Skip an intermittent issue tests/tail/overlay-headers (fails in this run but passes in the 'main' branch)
Skipping an intermittent issue tests/timeout/timeout (passes in this run but fails in the 'main' branch)