stty: verify tcsetattr applied all requested settings

[melihemik]

Fixes #10324

POSIX allows tcsetattr() to return success even when it could only partially apply the requested settings:

The tcsetattr() function shall return successfully if it was able to perform any of the requested actions, even if some of the requested actions could not be performed.

GNU stty handles this by calling tcgetattr() after every write and comparing the result to what was requested. If they differ it exits non-zero. We did not do this at all.

The fix adds termios_eq() which checks the four flag groups (input/output/control/local), all control characters, and both baud rates — the same fields GNU's eq_mode() covers. Platform-specific fields like line_discipline are left out on purpose because the kernel may normalise them independently of what we wrote.

Error message matches GNU:

standard input: unable to perform all requested operations

Four unit tests cover the new helper: identical structs, one flag differing, a control character differing, and a baud rate differing.

[codspeed-hq]

Merging this PR will degrade performance by 1.98%

⚠️ 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
✅ 343 untouched benchmarks
⏩ 46 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	df_with_path	573.7 µs	704.7 µs	-18.59%
⚡	Simulation	du_summarize_balanced_tree[(5, 4, 10)]	16.8 ms	16 ms	+4.77%
⚡	Simulation	du_max_depth_balanced_tree[(6, 4, 10)]	65.2 ms	62.5 ms	+4.33%
⚡	Simulation	complex_relative_date	330.2 µs	318.4 µs	+3.72%

Tip

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

---

_{Comparing melihemik:stty-verify-tcsetattr (f612337) with main (822aa83)}

Footnotes

1. 46 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. ↩

[collinfunk]

The fix adds termios_eq() which checks the four flag groups (input/output/control/local), all control characters, and both baud rates — the same fields GNU's eq_mode() covers.

/// Compare two `Termios` structs for equality the same way GNU's `eq_mode()` does:
/// input/output/control/local flags, all control characters, and both baud rates.
/// We deliberately skip any platform-specific fields (like `line_discipline`) that
/// the kernel may normalise on its own. 
fn termios_eq(a: &Termios, b: &Termios) -> bool {
    a.input_flags == b.input_flags
        && a.output_flags == b.output_flags
        && a.control_flags == b.control_flags
        && a.local_flags == b.local_flags
        && a.control_chars == b.control_chars
        && cfgetispeed(a) == cfgetispeed(b)
        && cfgetospeed(a) == cfgetospeed(b)
}

Please license your contributions under GPLv3+.