expand: improve the performances - 1.80 faster than GNU

[sylvestre]

No description provided.

[sylvestre]

$ hyperfine --warmup 3 "/usr/bin/expand moby64.txt"                     "./target/release/coreutils.prev expand moby64.txt" "./target/release/coreutils expand moby64.txt"
Benchmark 1: /usr/bin/expand moby64.txt
  Time (mean ± σ):     116.6 ms ±   4.3 ms    [User: 102.2 ms, System: 14.4 ms]
  Range (min … max):   111.2 ms … 130.6 ms    26 runs

Benchmark 2: ./target/release/coreutils.prev expand moby64.txt
  Time (mean ± σ):     584.1 ms ±   4.3 ms    [User: 482.2 ms, System: 101.8 ms]
  Range (min … max):   578.3 ms … 591.6 ms    10 runs

Benchmark 3: ./target/release/coreutils expand moby64.txt
  Time (mean ± σ):      65.0 ms ±   2.4 ms    [User: 52.2 ms, System: 12.7 ms]
  Range (min … max):    61.6 ms …  70.8 ms    44 runs

Summary
  ./target/release/coreutils expand moby64.txt ran
    1.79 ± 0.09 times faster than /usr/bin/expand moby64.txt
    8.99 ± 0.34 times faster than ./target/release/coreutils.prev expand moby64.txt

[github-actions]

GNU testsuite comparison:

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

[codspeed-hq]

CodSpeed Performance Report

Merging #8803 will improve performances by 17.52%

_{Comparing sylvestre:expand-perf (fb5f4d8) with main (f860610)}

Summary

⚡ 2 improvements
✅ 75 untouched
⏩ 73 skipped¹

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
⚡	expand_custom_tabstops[50000]	30.5 ms	25.9 ms	+17.52%
⚡	expand_many_lines[100000]	106.6 ms	100.9 ms	+5.67%

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

[cakebaker]

Wow, almost 9x faster than before, kudos!