Re: [bug-gawk] Behavior of fflush with SIGPIPE on stdout [PATCH]

[FERRIEUX Alexandre - IMT/OLN]

On 25/03/2017 14:41, Andrew J. Schorr wrote:
> On Sat, Mar 25, 2017 at 11:49:31AM +0100, FERRIEUX Alexandre - IMT/OLN wrote:
> >  On 25/03/2017 01:16, Andrew J. Schorr wrote:
> >  >Why is it exiting successfully
> >  >when it catches a SIGPIPE? As far as I can tell, that's not normal.
> >
> >
> >  Agreed. And while you're at it, a full emulation of
> >  exiting-with-default-SIGPIPE-handler would be even cleaner: exit
> >  status should end up as 141 = 128+EPIPE.
> >
> >
> >         cat /etc/services | head -1 ; echo $?
> >         # /etc/services:
> >         141
>
> Hmmm. I'm not sure, but I suspect this may be platform-specific.
> Actually, testing on Cygwin, I don't get any error at all:
>
> $ uname -a
> CYGWIN_NT-10.0 ti47 2.5.0(0.297/5/3) 2016-04-11 09:58 x86_64 Cygwin
>
> $ false | head -1; echo $?
> 1
>
> $ cat /etc/services | head -1 ; echo $?
> # Copyright (c) 1993-2004 Microsoft Corp.
> 0
>
> I don't know why that's the case.

You forgot "set -o pipefail".
Here is a variant that doesn't need it:

  (cat /etc/services;echo $?>&2) | head -1
  # /etc/services:
  141