test_tail::test_follow_with_pid fails on freebsd

[sylvestre]

2023-01-27T21:17:25.3459300Z failures:
2023-01-27T21:17:25.3527170Z 
2023-01-27T21:17:25.3541950Z ---- test_tail::test_follow_with_pid stdout ----
2023-01-27T21:17:25.3555640Z current_directory_resolved: 
2023-01-27T21:17:25.3572230Z run: /Users/runner/work/coreutils/coreutils/target/debug/coreutils tail -f --pid=5483 foobar.txt foobar2.txt
2023-01-27T21:17:25.3589780Z open: /tmp/.tmp7J1aby/foobar_follow_multiple.expected
2023-01-27T21:17:25.3601920Z write(append): /tmp/.tmp7J1aby/foobar2.txt
2023-01-27T21:17:25.3609060Z write(append): /tmp/.tmp7J1aby/foobar.txt
2023-01-27T21:17:25.3612710Z open: /tmp/.tmp7J1aby/foobar_follow_multiple_appended.expected
2023-01-27T21:17:25.3639510Z write(append): /tmp/.tmp7J1aby/foobar.txt
2023-01-27T21:17:25.3647600Z thread 'test_tail::test_follow_with_pid' panicked at 'Expected stdout to be empty, but it's:
2023-01-27T21:17:25.3655790Z should
2023-01-27T21:17:25.3687410Z be
2023-01-27T21:17:25.3689460Z ignored
2023-01-27T21:17:25.3705700Z ', tests/by-util/test_tail.rs:694:10
2023-01-27T21:17:25.3710730Z stack backtrace:
2023-01-27T21:17:25.3713630Z    0: rust_begin_unwind
2023-01-27T21:17:25.3727760Z              at /rustc/fc594f15669680fa70d255faec3ca3fb507c3405/library/std/src/panicking.rs:575:5
2023-01-27T21:17:25.3730800Z    1: core::panicking::panic_fmt
2023-01-27T21:17:25.3740340Z              at /rustc/fc594f15669680fa70d255faec3ca3fb507c3405/library/core/src/panicking.rs:64:14
2023-01-27T21:17:25.3746820Z    2: tests::common::util::CmdResult::no_stdout
2023-01-27T21:17:25.3749800Z              at ./tests/common/util.rs:430:9
2023-01-27T21:17:25.3750580Z    3: tests::test_tail::test_follow_with_pid
2023-01-27T21:17:25.3754630Z              at ./tests/by-util/test_tail.rs:689:5
2023-01-27T21:17:25.3758620Z    4: tests::test_tail::test_follow_with_pid::{{closure}}
2023-01-27T21:17:25.3764420Z              at ./tests/by-util/test_tail.rs:636:27
2023-01-27T21:17:25.3764750Z    5: core::ops::function::FnOnce::call_once
2023-01-27T21:17:25.3766410Z              at /rustc/fc594f15669680fa70d255faec3ca3fb507c3405/library/core/src/ops/function.rs:507:5
2023-01-27T21:17:25.3769420Z    6: core::ops::function::FnOnce::call_once
2023-01-27T21:17:25.3778600Z              at /rustc/fc594f15669680fa70d255faec3ca3fb507c3405/library/core/src/ops/function.rs:507:5
2023-01-27T21:17:25.3781080Z note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
2023-01-27T21:17:25.3781880Z 
2023-01-27T21:17:25.3783010Z 
2023-01-27T21:17:25.3790450Z failures:
2023-01-27T21:17:25.3794280Z     test_tail::test_follow_with_pid
2023-01-27T21:17:25.3795100Z 
2023-01-27T21:17:25.3795780Z test result: FAILED. 2363 passed; 1 failed; 38 ignored; 0 measured; 0 filtered out; finished in 164.95s

@Joining7943 rings a bell ? :)

[Joining7943]

rings a bell ? :)

Not really, to be honest :) Why?

However, I think I've seen this error before once, so I guess it's an intermittent failure? I couldn't reproduce this behavior when I ran this test a several times locally with freebsd installed in a virtual box but I'll try to figure it out.

[sylvestre]

because I saw your name in the blame and you recently touched a bunch of things :)

No worries if you don't know
We could just ignore this test on freebsd. wdyt ?

[Joining7943]

because I saw your name in the blame and you recently touched a bunch of things :)

I hope for the better but it's right, I changed the test to adjust it to use UChild and added additional assertions. The test structure itself hasn't changed.

No worries if you don't know

yeah I don't know, but since this test only fails on freebsd sometimes, I guess the delay after the kill should be raised. macos and freebsd are really slow in the ci and looking at the code in tail::follow::watch, it's not impossible that the appended content is printed after the 1 secs delay even without the followed process being alive.

We could just ignore this test on freebsd. wdyt ?

Ok. Let's ignore it on freebsd for now.