use LWP::Simple qw( get );

my $image = get("http://example.com/images/file.jpg");
defined($image)
   or die("Unable to fetch image\n");

my $image = GD::Image->new($image);
[download]

Looks right to me, but beware that this loads the whole image into memory. If this URL is being supplied by untrusted users you'll need to do something to prevent being crashed by someone pointing you to a gigantic image. Even trusted users can do this - I once had users uploading 200MB+ uncompressed images into Krang so they could include them on web-pages (shrunk via HTML width/height attributes, of course!).

Anybody know how you tell LWP to stop if it gets to X bytes? Some kind of read-callback, yes?

-sam

Can't do it with LWP::Simple (natch), but yes if you use $ua->get( $url, :content_cb => ... ) you can pass a callback which dies after the requisite number of bytes have been received.

#!/opt/local/bin/perl

use strict;
use LWP::UserAgent ();

my $ua = LWP::UserAgent->new();
$ua->env_proxy;

my $rcvd = 0;
my $buf = '';
my $resp = $ua->get( 'http://www.google.com/',
                     ':content_cb' => sub { $buf .= $_[0]; die "NO MOR
+\n" if ( $rcvd += length $_[0] ) > 512; });

print "content:\n", $buf, "\n\n";

print "headers:\n", $resp->as_string, "\n";
exit 0;

__END__
[download]

If the callback gets too much your response object will have an X-Died header (if you care if it did or not).

The cake is a lie.
The cake is a lie.
The cake is a lie.

But does that work for JPEGs as well as GIFs? I remember writing an image map program in TK and had to physically open a JPEG, and do a bit of unpacking magic on some of the first 9 bytes of the file to get the file dimensions. Granted this was years ago and maybe the module has been updated but as I recall, I tried that and it wouldn't handle the JPEGs.

Update: Realized the OP was asking about remote images not local.

---

Revolution. Today, 3 O'Clock. Meet behind the monkey bars.

I would love to change the world, but they won't give me the source code

You can have a look at WWW::Mechanize and $mech->find_image() who returns a WWW::Mechanize::Image object

Update:

#!/usr/bin/perl 
use strict;
use warnings;
use WWW::Mechanize;
use Data::Dumper;

my $url = q{http://perlmonks.org/};
my $m = WWW::Mechanize->new();
$m->get($url);
print Dumper $_ for @{ $m->images };
[download]

Output:

$VAR1 = bless( {
                 'width' => '468',
                 'base' => bless( do{\(my $o = 'http://perlmonks.org/'
+)}, 'URI::http' ),
                 'alt' => 'Beefy Boxes and Bandwidth Generously Provid
+ed by pair Networks',
                 'name' => undef,
                 'url' => 'http://promote.pair.com/i/pair-banner-curre
+nt.gif',
                 'tag' => 'img',
                 'height' => '60'
               }, 'WWW::Mechanize::Image' );

$VAR1 = bless( {
                 'width' => '74',
                 'base' => bless( do{\(my $o = 'http://perlmonks.org/'
+)}, 'URI::http' ),
                 'alt' => 'Ovid',
                 'name' => undef,
                 'url' => 'http://perlmonks.org/images/usermonkpics/ov
+idmonk.gif',
                 'tag' => 'img',
                 'height' => '91'
               }, 'WWW::Mechanize::Image' );
[download]

hth,
PooLpi

It's important to distinguish that as an HTML attribute for displaying the image and not as a property of the image itself. Have to have the image (its data) to read that.

I'm not clear on your objective.

Are you trying to avoid downloading the image (i.e., do want to get to its attributes without...or before you...download it?)

Or is it OK to download it and then get the attributes?