while (@found < $pics_to_find * 30) {
  last unless @found_galleries;  # exit the loop if there are no more
  my $gallery_content = get(pop @found_galleries);

  while ($gallery_content =~ /.../) {
    ...

    # why use "yes" and "no"?  why not just a true value or a false va
+lue?
    next if
      $limit_img_size eq "yes"
      and !some_size_constraint();

    push @found, ...;
  }
}
[download]

The problem is that you never update $pics_found. Replace $pics_found with @found_images.

I see that you tried to debug the problem by adding print statements. That's good. (We like monks that put in some effort, and it's what I would have done.) The problem with your print statements is that they are being buffered. Adding $| = 1 will reveal them.

I took the liberty of cleaning up your code:

# $| = 1;  # Uncomment when adding print statements for debugging.

# Get more than needed so we can randomly choose images later.
while (@found_galleries && @found_images < $pics_to_find * 30)
{
   # Remove one gallery link at a time until we're done.
   my $gal_url = pop(@found_galleries);
   my $gal = get($gal_url);

   # Extract the image URLs from the gallery page.
   my @image_urls = $gal =~ m#(http://images\.imagefap\.com/images/thu
+mb/\d+/\d+/\d+\.jpg)#g;

   # Remove the images which are too big.
   if ($limit_img_size eq "yes")
   {
      @image_urls = grep {
         my $image = get($_);
         my ($height, $width) = imgsize(\$image);
         $height <= $max_height && $width <= $max_width
      } @image_urls;
   }

   push @found_images, @image_urls;
}
[download]

Then again, maybe OPs code was working. Maybe he thought it was endless due not getting any prints back and because every image has to be loaded into memory for it to be sized. From ImageFap, that's one long wait because they are super slow.

Code might not be pretty but other than the loop variable not increasing, I think it should work the way it is. I'd try to turn on the buffer var and see what happens.

UPDATE: would it not be better for the server in terms of CPU and memory if the images were downloaded into a temp directory and sized locally rather than read from memory?

I don't understand what you are trying to say in the first two paragraphs. (Yes, the debug statements would have worked with $| = 1. Yes, it would have worked by replacing $pics_found with @found_images. Just like I said in my post.)

If your point is that you're taking exception to me cleaning up the code, keep in mind that good code is readable and maintainable code. The excess of redundancy and the poor choice of variable names in the original code affected both readability and maintainability.

would it not be better for the server in terms of CPU and memory if the images were downloaded into a temp directory and sized locally rather than read from memory?

The CPU usage would be (slightly) higher, since we'd have extra code to write to the disk and read from the disk.

Yes, the memory usage could be smaller (depending on how we write to and read from the disk) by something less than the size of one image.

And of course, it would be slower because we'd have to do all the work we currenlty do, plus more.

Hi. I have been playing with this all afternoon and I can't get the output to work.

It just doesn't size some of the images.

my $pics_found = $#found_images + 1;

while (@found_galleries && @found_images <= $pics_to_find * 20)
{
   # Remove one gallery link at a time until we're done.
   my $gal_url = pop(@found_galleries);
   my $gal = get($gal_url);
   print "gal = $gal_url<br>";

   # Extract the image URLs from the gallery page.
   my @image_urls = $gal =~ m#(http://images\.imagefap\.com/images/thu
+mb/\d+/\d+/\d+\.jpg)#g;
   foreach my $img (@image_urls) { $img =~ s/thumb/full/ig; }

   # Remove the images which are too big.
   if ($limit_img_size eq "yes")
   {
      @image_urls = grep {
         my $this_url = $_;
         my $image = get($this_url);
         my ($height, $width) = imgsize(\$image);
         print "IMG $this_url is $height and $width<br>";
         $height <= $max_height && $width <= $max_width
      } @image_urls;
   }

   push @found_images, @image_urls;
}
[download]

The output below is generally what I get. It will put the size down for a few of them but only rarely.

gal = http://www.imagefap.com/gallery.php?gid=158098
IMG http://images.imagefap.com/images/full/5/190/190037365.jpg is and
IMG http://images.imagefap.com/images/full/6/804/804735625.jpg is and
IMG http://images.imagefap.com/images/full/7/478/478782005.jpg is and
IMG http://images.imagefap.com/images/full/10/199/1994599748.jpg is an
+d
IMG http://images.imagefap.com/images/full/4/115/1151974238.jpg is and
[download]

I don't suppose you or anyone else can see what's going wrong?

It works for me, or rather, the following does:

use strict;
use warnings;

use Data::Dumper qw( Dumper );
use Image::Size  qw( imgsize );
use LWP::Simple  qw( get );

my $this_url = "http://images.imagefap.com/images/full/5/190/190037365
+.jpg";
my $image = get($this_url);
my ($height, $width) = imgsize(\$image);
print "IMG $this_url is $height and $width\n";

__END__
IMG http://images.imagefap.com/images/full/5/190/190037365.jpg is 507 
+and 760
[download]

Make sure $image is defined. Maybe an error occured during the download?

print(defined($image) ? "Ok." : "Error!", "<br>");
[download]

Oh! And make sure your version of Image::Size accepts images passed as \$image.

PS -
foreach my $img (@image_urls) { $img =~ s/thumb/full/ig; }
can be written as
s/thumb/full/i foreach @image_urls;