Redirecting STDOUT

tame1 has asked for the wisdom of the Perl Monks concerning the following question:

I find myself in need of a way of redirecting/capturing STDOUT, but from within and program instead of from the command line.

In detail, here is the problem. We use a home grown templating system in which the template holds some perl and lots of html, with a bunch of $href->{somekey} entries.
That template is slurped up, and then eval'ed (with $href having been created/filled before the slurp and eval).

The slurped template can be edited by non-programmers, so for safety we 'use Safe'. Here is the code so far:

  my $code;
  my $compartment = new Safe; 
  $compartment->share('$href');
  $code = $compartment->reval($$file);
  if ($@) {
    _sAbort( $r, "Restricted eval failed! Unable to compile format of 
+type $type:", $@ );
    return DONE;
  }
[download]

The real problem now comes up. ANY use/call of $code->($href) will activate the print statements inside the slurped file. This was fine with the following:

$r->send_http_header('text/html');
$code->($href);
return DONE;
[download]

But now I am trying to instead grab the output of $code and stick it as text into an email. (You guessed it - I'm trying to email generated web pages - HTML email, yuck!).

Surfing through the various books, I find no simple way to redirect STDOUT to a variable for capture. The following code failed miserably:

my $open_err;
my $data;
open(OUTPUT, "$code->($href) |") or $open_err = $!; 
if ($open_err) {
    _sAbort($r,"Cannot open a pipe",$!);
    return DONE;
}
local $/ = undef;
$data = <OUTPUT>;
close(OUTPUT);
[download]

Any of you geniouses know offhand how I can get this darned HTML output from the code ref $code into a simple variable so i can give it to MIME::Lite as data? Much gracious bowing and licking of boots are offered to the one who can solve this for me!

What does this little button do . .<Click>; "USER HAS SIGNED OFF FOR THE DAY"

Comment on Redirecting STDOUT Select or Download Code

Replies are listed 'Best First'.
Re: Redirecting STDOUT by SarahM (Monk) on Jun 14, 2002 at 13:54 UTC
Here is a trick I've used before. I know it works on windows...but I haven't tried it on unix. You'll need to download IO::Scalar, it comes with IO-stringy. `use IO::Scalar; my $output; tie (STDOUT, 'IO::Scalar', \$output); # Use this to capture STDOUT print "testout\n"; untie STDOUT; # Use this to return STDOUT to normal print "my output='$output'";` [download] I hope this works for you.	[reply] [d/l]
Re: Re: Redirecting STDOUT by mt2k (Hermit) on Jun 15, 2002 at 04:20 UTC
I never thought of tieing the variable... this is how I use IO::Scalar when I wish to redirect output. It's actually more useful for when you just need to have output to a variable rather than to some file handle: `use IO::Scalar; my ( $STD1 $STD2 ); $STD1 = new IO::Scalar \$STD2; select $STD1; print <<DONE; blah blah blah All this is being appended to the $STD2 variable So some more text here and we are... DONE open FILE, ">test.txt"; print FILE $STD2; close FILE;` [download] Yes, odd example since you could have just used: `open FILE, ">test.txt"; select FILE; print <<DONE; all that stuff again. I am DONE` [download] But you get the point...	[reply] [d/l] [select]
Re: Redirecting STDOUT by Abigail-II (Bishop) on Jun 14, 2002 at 14:44 UTC
If you are on a Unix like system, you can use `open my $fh => "\|-"` and `open my $fh => "-\|"` as described in `perlipc`. Also very useful for post processing output (or preprocessing input) without changing your existing program. Abigail	[reply] [d/l] [select]
Re: Redirecting STDOUT by Dog and Pony (Priest) on Jun 14, 2002 at 13:50 UTC
I am not sure if this will help you, but you can find several (mine and some others) takes on capturing STDOUT at STDOUT::Capture - manipulate STDOUT. You have moved into a dark place. It is pitch black. You are likely to be eaten by a grue.	[reply]
(RhetTbull) Re: Redirecting STDOUT by RhetTbull (Curate) on Jun 14, 2002 at 16:56 UTC
This question comes up a lot... I know I've answered it at least eight times. As always, my recommended solution is to use Filter::Handle. It does a great job, requires minimum changes to existing code, and works with file handles other than STDOUT/STDERR.	[reply]
Re: (RhetTbull) Re: Redirecting STDOUT by Anonymous Monk on Jun 14, 2002 at 22:04 UTC
And, as pointed out in its documentation, there are things it cannot catch. The most common of which are system calls.	[reply]
Re: Redirecting STDOUT by Anonymous Monk on Jun 14, 2002 at 14:21 UTC
I know how to redirect STDOUT to a file open (STDOUT, ">>myfile.log"); Works just fine for STDERR also! To redirect into a variable, perhaps you can try `@LINES = <STDOUT>;` [download]	[reply] [d/l]
Re: Redirecting STDOUT by BUU (Prior) on Jun 15, 2002 at 18:43 UTC
Is it to late to point out that there are tons of very nice templating engines already available, from html::template to ePerl? For a quick overview, you can look at "http://www.perl.com/pub/a/2001/08/21/templating.html", which describes many of the common engines out there.	[reply]


Pathologically Eclectic Rubbish Lister
	PerlMonks