Documentation is a bit rough

I've annotated a short test script that was kicking around in my test directory. Maybe it'll clarify things a little?

#! perl -slw
use strict;
use Win32::Socketpair qw[ winopen2_5 ];

## Open a pipe to perl.exe -c
## Retrieving the pid and the socket handle
my( $pid, $sock ) = winopen2_5( 'perl.exe', '-c' );

## Print a short script to it for syntax checking
print $sock <<'EOP';
my $x = 365.25**3;
my $y = qx[ perl -c -e"say 'boo'"; ];
my $z = sub {
    print $sock <<;
    my $x = 365.25**3;
    my $y = qx[ perl -c -e"say 'boo'"; ];
    my $z = sub {
        print $sock <<;
        my $x = 365.25**3;
        my $y = qx[ perl -c -e"say 'boo'"; ];
        my $z = sub {
            1;
        };
    };
};
EOP

## Tell the socket that we have finsihed writing to it
## Has the effect of allowing the pelr process to know 
## that no more data is coming, so it can process the script
shutdown $sock, 1;

## Read back all the output from the perl process 
## -- stdout & stderr -- and display on terminal
print while <$sock>;

## Close the socket
close $sock;

## Start another copy of perl with a one liner argument
## that outputs to both stdout and stderr
( $pid, $sock ) = winopen2_5( 'perl.exe', q[-E"say 'hello'; warn; die;
+"] );

## Tell the socket we've finished writing
shutdown $sock, 1;

## print teh retrieved output to the terminal
print while <$sock>;

## and done
close $sock;

## The output produced is below
__END__
C:\test>winopen2_5
- syntax OK

Warning: something's wrong at -e line 1.

Died at -e line 1.

hello
[download]

Now that's quite a bit different and seems to make more sense. I am now able to get the output I desire but there appears to be some random characters that are printed with the first line. I assume that it'll wait for the process to complete before the shutdown call is made? Or is waitpid needed?

Also, I was checking $fh against $out and $err before (see OP), I don't really seem to have that option now. What can I do to compensate or should I not bother?

sub sys {
my $self = shift;
my $app  = shift;
more();
TRACE "sys > $app @_";
my($pid, $sock) = Win32::Socketpair::winopen2_5($app, @_) or LOGDIE "f
+ailed to open $app: @_";
shutdown $sock, 1;
my ($sel, $buf);
$sel = new IO::Select($sock);
SYSLOOP: while(my @ready = $sel->can_read) {
    foreach my $fh (@ready) {
        my $line = <$fh>;
        if(not defined $line) {
            $sel->remove($fh);
            next;
        }
        $buf .= $line;
    }
}
waitpid($pid, 0);
close $sock;
less();
return $buf;
}
[download]

Or is waitpid needed?

No, though it won't harm. On Windows, when a process terminates the OS stores the return code, so it can be retrieved, and kills the process. Unlike *nix where processes hang around as zombies until something asks them for their return code.

So on Windows all waitpit does -- assuming the process has actually already completed, which it will have by the time you've read all its output -- is just request the rc from the OS.

Also, I was checking $fh against $out and $err before (see OP), I don't really seem to have that option now. What can I do to compensate or should I not bother?

This call doesn't allow you to differentiate between that output that came from stdout and that from stderr -- hence the open2.5.

Whilst it would be possible to do a full open3 compatible call, I've never needed it.

---

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.