Tee resets exit status is always 0

Question

I have a short script like this:

#!/bin/bash
<some_process> | tee -a /tmp/some.log  &
wait $(pidof <some_process_name>)
echo $?

The result is always 0, irrespective of the exit status of some_process.

I know PIPESTATUS can be used here, but why does tee break wait?

You're checking the exit status of the last command which is the `wait` command. So if it's successfully finished, hence the exit status will be 0. — Avihoo Mamka, Mar 07 '16 at 11:33
@AvihooMamka No, `wait` should return the exit status of the specified pid. — Yam Marcovic, Mar 07 '16 at 11:36
@YamMarcovic from the [docs](http://linux.die.net/man/2/wait) - Return Value: wait(): on success, returns the process ID of the terminated child; on error, -1 is returned. — Avihoo Mamka, Mar 07 '16 at 11:40
@AvihooMamka You're looking at the docs of the system call, not the bash builtin..... — Yam Marcovic, Mar 07 '16 at 11:41
OK, from other [docs](http://wiki.bash-hackers.org/commands/builtin/wait): Return status: The return status is the return status of the job waited for, or 0 - waited for all jobs in shell's job list. 1 - the given ID is not a valid job or process ID — Avihoo Mamka, Mar 07 '16 at 11:43
I didn't say that. I said that the return status is 0 if "waited for all jobs in shell's job list". — Avihoo Mamka, Mar 07 '16 at 11:45
@AvihooMamka But is that the case here? He waits on a specific pid. — Yam Marcovic, Mar 07 '16 at 11:45
@YamMarcovic Yes but the process is still attached to the shell's job list. He should `disown` the job. — Avihoo Mamka, Mar 07 '16 at 11:49
He would. The process is being moved to the background but still attached to the shell's job list. — Avihoo Mamka, Mar 07 '16 at 11:54
@hek2mgl Please open this question. It's not a duplicate. He says he knows about PIPESTATUS, but is asking about combining it with `wait`. — Yam Marcovic, Mar 07 '16 at 12:07

Yam Marcovic · Accepted Answer · 2016-03-07T12:38:29.900

Well, this is something that, for some peculiar reason, the docs don't mention. The code, however, does:

int wait_for (pid) { /*...*/
/* If this child is part of a job, then we are really waiting for the
job to finish. Otherwise, we are waiting for the child to finish. [...] */

if (job == NO_JOB)
  job = find_job (pid, 0, NULL);

So it's actually waiting for the whole job, which, as we know, normally yields the exit status of the last command in the chain.

To make matters worse, $PIPESTATUS can only be used with the last foreground command.

You can, however, utilize $PIPESTATUS in a subshell job, like this:

(<some_process> | tee -a /tmp/some.log; exit ${PIPESTATUS[0]}) &
# somewhere down the line:
wait %%<some_process>

hek2mgl · Answer 2 · 2016-03-07T12:32:31.923

1

The trick here is to use $PIPESTATUS but also wait -n:

Check this example:

#!/bin/bash

# start background task
(sleep 5 | false | tee -a /tmp/some.log ; exit ${PIPESTATUS[1]}) &

# foreground code comes here
echo "foo"

# wait for background task to finish
wait -n
echo $?

edited Mar 07 '16 at 12:32

answered Mar 07 '16 at 12:21

hek2mgl

152,036
28
249
266

Good trick. I think you don't even need `-n` though, you can just give a specific jobspec. – Yam Marcovic Mar 07 '16 at 12:22
@YamMarcovic Looks like without `-n`, `wait` would always return `0`. I admit the `help wait` does not say that. Let me investigate that. – hek2mgl Mar 07 '16 at 12:22
Actually, this will block until the first job finishes. So I'm not sure this is useful. – Yam Marcovic Mar 07 '16 at 12:25
`wait` is supposed to block. Am I missing something? Sure, in a real world program between the background task and the wait command other lines of code would appear. – hek2mgl Mar 07 '16 at 12:26
It will block before wait. See my (edited) answer for a combination of your trick and the use of a subshell to actually get a nice outcome. – Yam Marcovic Mar 07 '16 at 12:28
Oh sorry, missed the subshell here. Have it in my test!! :) – hek2mgl Mar 07 '16 at 12:30
1

:) Also, you're losing the exit status now, but just returning 0/1. When in a subshell, you can simply exit with PIPESTATUS[N] (as in my answer =]) – Yam Marcovic Mar 07 '16 at 12:31
1

Combined forces, eh. :) – Yam Marcovic Mar 07 '16 at 12:33

score 0 · Answer 3 · answered Mar 07 '16 at 12:41

The reason you always get an exit status of 0 is that the exit status returned is that of the last command in the pipeline, which is tee. By using a pipe, you eliminate the normal detection of exit status of <some_command>.

From the bash man page:

The return status of a pipeline is the exit status of the last command, unless the pipefail option is enabled. If pipefail is enabled, the pipeline's return status is the value of the last (rightmost) command to exit with a non-zero status, or zero if all commands exit successfully.

So... The following might be what you want:

#!/usr/bin/env bash

set -o pipefail
<some_command> | tee -a /tmp/some.log  &
wait %1
echo $?

Tee resets exit status is always 0

3 Answers3