#!/usr/bin/perl
use warnings;
use strict;

#
# warning: untested code.
#

package main;
use WWW::Mechanize;
    
my $mechanize = WWW::Mechanize->new(autocheck => 1); 
    
# define useragent
$agent = "Mozilla/5.0";
$mechanize->agent($useragent);

# get the url
#
$mechanize->get("http://www.site.com/images/");
#
# get the images
#
my @images = $mechanize->images();
foreach my $img (@images) {
    my $time = time();
    my $filename = $time . ".jpg";
    my $mech2 = WWW::Mechanize->new(autocheck => 1); 
#
# Save the images
# Use the fact that WWW:Mechanize objects are 
# overloaded LWP::UserAgent objects.
#
# please see: 
# http://search.cpan.org/dist/WWW-Mechanize/lib/WWW/Mechanize.pm
# and look for $mech->get
#
    my $mech2->get($img->url(), ":content_file" => $filename );
}
[download]

Ok thanks, $1 showed the correct URL, when i printed it. What I forgot to mention, is that the file size of the - with Perl - downloaded file is 0kb. When i show the image (on my hard drive) in my browser it return the path of the file. And what do you exactly mean by "checksum (md5/sha1)"? Thanks, cafaro

You need the full url for $ua->mirror(). If the urls in the HTML source are relative, your regex won't create a full url to mirror.

You should probably use WWW::Mechanize's find_image() or find_all_images() method:

for my $img ($www->find_all_images()) {
  $www->mirror($img->url_abs());
}
[download]

update: you probably also don't need both an LWP::UserAgent and a WWW::Mechanize object, since WWW::Mechanize is a subclass of LWP::UserAgent. In fact, chances are, it will work better with just one WWW::Mechanize object.
"What should it profit a man, if he should win a flame war, yet lose his cool?"

Linux is not an image viewer - how exactly did you try to open it? What's the output of file $image.jpg (use a real image name)?

My guess is that your regex doesn't work as you expect it to, $1 contains an invalid URL and you get an error page mirrored.

Did you try to print the matched URLs to STDOUT?

And try this regex:

$content =~ m{<tr><td><img src="([^"]+)" alt="php-image" /></td></tr>}

Make sure the image is in fact an image, and not the garbled binary of some cgi wrapper. Open it with the original browser that you viewed it with, if it works you should know what to do. If it doesn't checksum (md5/sha1) a copy you saved from your browser, and then a copy that you retrieved with perl.

My bet is they are off.

The URL you cite is only possible if the server is set to provide a script of some sort (you seem to have discovered it's a PHP script already) when the directory http://www.site.com/images/ is requested, and the ?id=345435 part is passed as an argument to the script.

I don't think it's possible to enumerate the directory in this case, you need to know the ID of every image you want -- and the server may be set up so that the ID is different every time that image is to be requested. It's a sneaky way to attempt to prevent people from grabbing copyrighted pictures off a web site or linking to them from their own sites.