Question about pipes and exit codes

[William Atkinson]

Jason,

Thanks for the suggestion and response.
I originally tried open2 (haven't done open3 yet) but only see sigpipe when
the program at the "far" end of the pipe dies, regardless of the cause of
death.  I started on chapter 16 of the cookbook this morning and am at p.
562, so I'd say we are definitely in the same ballpark.  I'd also say I'm in
a bit deep too.  

The next thing that I'm going to try is fussing with pipe and fork to get a
sort of pipe loop.  If I can do that successfully, I can catch the signal
that the child died from.

Chip

> -----Original Message-----
> From: Jason Van Slyke [mailto:jvanslyk at matchlogic.com]
> Sent: Wednesday, November 17, 1999 9:13 AM
> To: 'boulder-pm-list at happyfunball.pm.org'
> Subject: RE: Question about pipes and exit codes
> 
> 
> Chip,
> 
> First up, you're way over my head.  But I found this 
> interesting so I dug
> out a couple, 3 or 4 books.
> 
> Do you have a Perl Cookbook available?  If so, see p567-571.
> 
> If not, this _might_ help:
> 
> eval {
> 	open2($readme, $writeme, @program_and_arguments) ;
> } ;
> if ($@) {
> 	if ($@ =~ /^open2/) {
> 		warn "open2 failed: $!\n$@\n" ;
> 		return ;
> 	}
> 	die ;		#reraise unforeseen exception
> }
> 
> Let me know if this is playing in the same ball park as you are.
> 
> Jason
> 
> -----Original Message-----
> From: William Atkinson [mailto:CWA at DISC.com]
> Sent: Wednesday, November 17, 1999 8:46 AM
> To: 'boulder-pm-list at pm.org'
> Subject: Question about pipes and exit codes
> 
> 
> Greetings,
> 
> Forgive and politely correct me if this is the wrong list to post to.
> 
> I'm having trouble getting the exit status of a process at 
> the end of a pipe
> created by open2.  It seems that all I get is notification 
> that the pipe
> broke.  
> 
> One idea that we came up with was to have a shell script run 
> the desired
> program, catch its exit status and return it as the last string.  The
> problem with this is that I was hoping for a purely perl 
> based solution due
> to cross platform portability issues.
> 
> Any help would be appreciated.
> 
> Thanks.
> 
> Chip Atkinson
> chip at disc.com
>