Unexpand use buffered reads + tests

[cerdelen]

New PR to fix issues after revert of #10798

Fix: #10698

[sylvestre]

do we need that many tests ? thanks

[cerdelen]

Every test is a different edge case around the chunking process. I thought rather have a few more than a few too little.

[github-actions]

GNU testsuite comparison:

Congrats! The gnu test tests/tail/tail-n0f is now passing!

[sylvestre]

yeah my point was that there is some duplications in the tests and they might be merged with a loop with different args / expectations

[codspeed-hq]

Merging this PR will improve performance by 58.43%

⚡ 2 improved benchmarks
✅ 282 untouched benchmarks
⏩ 38 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	unexpand_many_lines[100000]	300.6 ms	189.7 ms	+58.4%
⚡	Simulation	unexpand_large_file[10]	630.1 ms	397.7 ms	+58.43%

---

_{Comparing cerdelen:unexpand_buffer (d4601f0) with main (194d980)}

Footnotes

1. 38 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. ↩