tty: add missing newline to output

[ChrisX101010]

Fixes #10239

[ChrisDryden]

(/bin/tty >&- 2>/dev/null); echo $?

This command shows what happens when theres an error in writing to stdout, the error code that GNU is expecting is 3, but by adding that ? it will make it handled and return a 1. You can get around this by doing:

Ok(name) => stdout
    .write_all_os(name.as_os_str())
    .and_then(|_| writeln!(stdout)),

[oech3]

target/debug/tty > /dev/full silently returns 3. But GNU have tty: write error: No space left on device.

[ChrisX101010]

@ChrisDryden can you check now, sir? thank you!

[oech3]

Would you add test to check newline (and exit code of >/dev/full if possible ) to
https://github.com/uutils/coreutils/blob/main/tests/by-util/test_tty.rs ? Thankyou.

[github-actions]

GNU testsuite comparison:

Note: The gnu test tests/basenc/bounded-memory is now being skipped but was previously passing.

[ChrisX101010]

@oech3 Hi, sir! Please check now and let me know. Thank you! Have a nice weekend. :)

[ChrisDryden]

Why did you change this?

[ChrisX101010]

@ChrisDryden I removed it because I thought it was redundant (flush was already called in line 37), but I've now restored it to match the original pattern.

[oech3]

I guess we can extract it from OS provided err message.
(Also please cargo fmt)

[ChrisX101010]

I changed it to print the error message before exiting (as requested in the previous feedback about matching GNU tty's 'write error' output).

However, I notice I removed the separate stdout.flush() check. The flush is still being called in the Ok(name) branch (line 37), but I can restore the separate check if you prefer. Should I keep both checks separately?

[codspeed-hq]

CodSpeed Performance Report

Merging this PR will degrade performance by 27.39%

_{Comparing ChrisX101010:fix-tty-missing-newline (2c92d38) with main (a1545ef)}

Summary

⚡ 1 improved benchmark
❌ 7 regressed benchmarks
✅ 274 untouched benchmarks
⏩ 38 skipped benchmarks¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Memory	sort_numeric[500000]	75.5 MB	79.2 MB	-4.66%
❌	Memory	sort_ascii_only[500000]	22.2 MB	28.3 MB	-21.77%
❌	Memory	sort_unique_locale[500000]	33.6 MB	39.8 MB	-15.51%
❌	Memory	sort_mixed_data[500000]	22.9 MB	27.1 MB	-15.71%
❌	Memory	sort_key_field[500000]	47.8 MB	51.8 MB	-7.62%
❌	Memory	sort_accented_data[500000]	22.1 MB	28.3 MB	-21.79%
❌	Memory	sort_long_line[160000]	712.6 KB	981.4 KB	-27.39%
⚡	Memory	dd_copy_default	130.8 KB	126.9 KB	+3.08%

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. ↩

[sylvestre]

many jobs are failing

[ChrisX101010]

many jobs are failing

I ran tests locally, and everything worked for me. I don't know why these specific jobs keep failing when I fixed the new line issue. There is no other workaround that I know of, GNU panics if something is wrong but it didn't produce an output so I made custom test like OP mentioned and I don't understand why is it failing now.. Can you suggest fix to this?

[oech3]

I guess you wrote the test writing to /dev/full based on @naoNao89 wrote one.
When I tried to do same things at different PR to remove panic, it failed.

I guess it is generic test framefork's issue.

[ChrisX101010]

I guess you wrote the test writing to /dev/full based on @naoNao89 wrote one. When I tried to do same things at different PR to remove panic, it failed.

I guess it is generic test framefork's issue.

Thank you, sir. If it's okay on your end, can we close this issue for now and this other issue can be a separate PR?

[oech3]

Please keep open and add ignore to the test.

[ChrisX101010]

Please keep open and add ignore to the test.

Understood, sir. Will follow your advice, thank you for the suggestion.