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

Hello all! I just registered here because I don't find any solution to my PERL problem.

So I want to strip certain tags from HTML in my PERL based Content Management System called Imperia.

I installed HTML::Restrict on the server. It works just fine in a little PERL test program I wrote:

#! /usr/local/bin/perl

use strict;
use HTML::Restrict;

my $value = qq|<html>
<head>
<title>Dies ist der Titel</title>
</head>
<body>
<p>Hier kommt der Inhalt</p>
</body>
</html>
|;

my $hr = HTML::Restrict->new();

my $value = $hr->process($value);

print "Ergebnis: ".$value."\n\n";

[download]

So now I want to use the module in my CMS. There is an interface called "transform". So I wrote this transform plugin:

#! /bin/false
# $Id: Null.pm,v 1.7.2.2 2009/03/26 11:15:04 guido Exp $ 
# vim: set autoindent shiftwidth=4 tabstop=8:

# Imperia AG is the sole owner and producer of its software "Imperia".
+ For
# our software license and copyright information please refer to: Lice
+nse.txt
# Copyright (C) 1995-2012 Imperia AG.  All rights reserved.

package Dynamic::Transform::zim_iwe_filtern;

use strict;
use warnings;
use HTML::Restrict;
use Carp;

sub new {
    my ($proto, $config) = @_;
    my $class = ref($proto) || $proto;

    bless {}, $class;
}

sub transform {
    my ($self, $metainfo) = @_;

  ## fetch all variable names
  my @keylist = sort $metainfo->keys();
  my $count = 0;
  foreach my $key (@keylist) {

    ## only change certian variables
    next unless (
      $key =~ m/^eingabe_/ or
      $key =~ m/^ff_confirm_/ or
      $key =~ m/^ff_successful_/ or
      $key =~ m/^map_point/ or
      $key =~ m/^map_text_/ or
      $key =~ m/^map_transcript_/ or
      $key =~ m/^tube_text_/
    );

    ## test value
    my $value = qq|<html>
  <head>
    <title>Dies ist der Titel</title>
  </head>
  <body>
    <p>Hier kommt der Inhalt</p>
  </body>
</html>
|;

    my $hr = HTML::Restrict->new() or die "Error at HTML::Restrict->ne
+w: $!";
    $value = $hr->process($value) or die "Error at process: $!";
    $metainfo->setValues($key."_DEBUG", $value);
  }
}


1;

[download]

But in this environment the command $value = $hr->process($value) won't work. The result on the web interface of the cms is as follows: "Error at process: at /cms/ews-2000/site/modules/core/Dynamic/Transform/zim_iwe_filtern.pm line 55, <DEFAULTDATEI> line 102."

This means that something goes wrong but the error variable $! isn't filled. The httpd error log shows no error either.

I suspect that the module isn't fit to be used inside a package or I just invoke it wrong.

Can anyone help me?

Kind regards,
Tobias from Germany

Replies are listed 'Best First'.
Re: HTML::Restrict works on the command line, but not the web server
by Corion (Patriarch) on Apr 26, 2016 at 11:53 UTC

    Note that one cause for the error might be your addition of ... or die "Error at process: $!". You don't do this in your test script that works from the command line.

    To get a better description of the error (if there is an error raised), I would wrap the offending line in an eval {} block:

    my $ok = eval {
    $value = $hr->process($value);
    1;
};
if( ! $ok ) {
    die $@; # reraise our error
};

    [download]

    If that still leaves an empty value in $value, then maybe the problem is in HTML::Restrict. Maybe you need to pass it some rules as to what is allowed.

[reply]
[d/l]
[select]

      Thanks for the immediate ansswer!

      I have changed the program like you suggested, but now no error shows up and the $value always comes out empty.

      If no rules are passed, all the HTML is filtered in my test script.

[reply]

        Then maybe something is wrong with either how HTML::Restrict works or with how you inspect the results.

        Have you made sure that Perl loads the correct versions of HTML::Restrict in both cases?

        print $INC{'HTML/Restrict.pm'};
print HTML::Restrict->VERSION;

        [download]

        Are you certain that your loop even gets entered? You loop over various (submitted form) values, but maybe they are not even there? Maybe retest with a more minimal test package:

        package Dynamic::Transform::zim_iwe_filtern;

use strict;
use warnings;
use HTML::Restrict;
use Carp;

sub new {
    my ($proto, $config) = @_;
    my $class = ref($proto) || $proto;

    bless {}, $class;
}

sub transform {
    my ($self, $metainfo) = @_;

    ## test value
    my $value = qq|<html>
  <head>
    <title>Dies ist der Titel</title>
  </head>
  <body>
    <p>Hier kommt der Inhalt</p>
  </body>
</html>
|;

    my $hr = HTML::Restrict->new() or die "Error at HTML::Restrict->ne
+w: $!";
    $value = $hr->process($value) or die "Error at process: $!";
    #$metainfo->setValues($key."_DEBUG", $value);
warn "Have [$value] after HTML::Restrict";
}

1;

        [download]

        Also, how certain are you that your CMS actually has reloaded the modified version of your code? The safest way is to restart the HTTP server, but also consider having a version number in your package that you output into a log file to make certain that the code you're looking at is also the code you are running.

[reply]
[d/l]
[select]

Back to Seekers of Perl Wisdom