expand, unexpand: flush output after each line

[r3wretrhy]

When stdin is read line-at-a-time (interactive terminal, or any producer that pauses between lines), expand / unexpand held converted output in a BufWriter and only flushed on exit. An interactive session therefore showed no converted line until the user quit, while GNU shows each line immediately.

Like cut, skip BufWriter when stdout is a terminal. On the buffered path, flush once per input chunk that completed a line (before the next read), so line-at-a-time pipes also see output promptly without a write per line on large inputs.

$ expand -t3
        a          # typed tab+a, Enter
   a               # GNU shows this right away; we used to wait until quit

Added test_stdin_tty_flushes_each_line to test_expand.rs and test_unexpand.rs.

Closes #11091

[codspeed]

Merging this PR will regress 1 benchmark

⚠️ 48 benchmarks spent significant time in system calls

System calls cannot be consistently instrumented, so they are not included in the measure, which understates the real cost. Please switch to the Walltime instrument to accurately measure system calls.

Measurement and system calls

• cksum_multiple_files
• dd_copy_64k_blocks
• cksum_default
• cksum_crc32b
• du_wide_tree[(5000, 500)]
• cut_fields_newline_delim
• dd_copy_8k_blocks
• cksum_sysv
• hostname_ip_lookup[100000]
• du_all_wide_tree[(5000, 500)]
• …and 38 more

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 3 improved benchmarks
❌ 1 regressed benchmark
✅ 76 untouched benchmarks
⏩ 391 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	factor_three_39_bit_primes	715.7 ms	795.8 ms	-10.06%
⚡	factor_five_38_bit_primes	2 s	1.8 s	+13.46%
⚡	expand_many_lines[100000]	115.9 ms	106 ms	+9.31%
⚡	expand_custom_tabstops[50000]	31.6 ms	29.7 ms	+6.51%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing r3wretrhy:expand-unexpand-stdin-flush (a56ecf7) with main (98194d4)²}

Footnotes

1. 391 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. ↩

2. No successful run was found on main (9f52800) during the generation of this report, so 98194d4 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[oech3]

Can we avoid large performance drop?

[github-actions]

GNU testsuite comparison:

Congrats! The gnu test tests/touch/now-owned-by-other is no longer failing!

[r3wretrhy]

@oech3 The perf drop is fixed in 0a02252, it only flushes when stdin is a pipe/tty now. CodSpeed shows expand_many_lines +9% and expand_custom_tabstops +6.5% faster. The factor regression looks unrelated since this PR doesn't touch factor. Also enabled rustix for all unix targets so it builds on macOS/BSD.