Re: Does head util cause SIGPIPE?

[Ray Satiro]

On 10/25/2019 6:37 PM, Assaf Gordon wrote:
> On 2019-10-25 1:56 a.m., Ray Satiro wrote:
> > Recently I tracked down a delay in some scripts to this line:
> >
> > find / -name filename* 2>/dev/null | head -n 1
> [...]
> > owner@ubuntu1604-x64-vm:~$ ( trap '' pipe; find / -name initrd* 
> > 2>/dev/null | strace -e 'trace=!all' head -n 1)
> > /initrd.img
> > +++ exited with 0 +++
> > (few seconds wait)
>
> In your case,
> I can guess that there is only a single file matching your predicate 
> 'initrd*'.
> The 'head' indeed terminates, and the pipe is closed.
> But if 'find' doesn't find any more matching files, it doesn't
> try to print anything more, and SIGPIPE is never raised.
>
> Note the manual page of pipe(7) says:
> ?????? "If all file descriptors referring to?? the?? read
> ???????? end?? of a pipe have been closed, then a write(2)
> ???????? will cause a SIGPIPE signal to be generated for the
> ???????? calling process."
>
> So if no further files were found, 'find' continues (slowly scanning
> the disk) until it finishes. 

Thanks for all the help guys. The explanation quoted above is indeed 
what is happening: find is only killed by SIGPIPE after it writes again, 
which in this case it does after it finds the second result.

I probably caused some confusion with my example because I made a 
mistake. For the purposes of the example I was trying to demonstrate I 
had reset the sigpipe trap, but actually I ignored it (oops). I wrote 
trap '' pipe which *ignores* sigpipe (in other words find will keep 
running after SIGPIPE), but I meant trap - pipe to reset it. Sorry.

I can confirm:

First window:
( trap - pipe; find / -name initrd* 2>/dev/null | head -n 1)

And in second window:
sudo strace -ff -e trace=write?? -e write=1 -p `pgrep find`

Then:

write(1, "/usr/src/linux-headers-4.4.0-34/"..., 127) = -1 EPIPE (Broken 
pipe)
--- SIGPIPE {si_signo=SIGPIPE, si_code=SI_USER, si_pid=7252, 
si_uid=1000} ---
+++ killed by SIGPIPE +++

So that's what's happened, killed by SIGPIPE but only after the next 
write. To remedy I am using the find ... -print -quit instead of head -n 
1 to get just the first result , and I will use Stephane's script if I 
need to get the first n results.

* Note I am not subscribed to the list, I can't get mail from the list 
for some reason (probably yahoo), so if a reply is necessary I will need 
a CC to see it.