update: You only have to do this if HTML::SimpleLinkExtor doesn't do it for you -- and it apparently will do if for you if there is a <BASE HREF> tag in the page that you fetch. For that matter, when you happen to fetch a page that has no such tag, you could just put one in before passing the page to LinkExtor:

s{(</head>)}{<base href="$current_docroot">$1}i;
[download]

I kinda thought you were already doing that (or trying to) in the code you posted -- but maybe you haven't had a chance to test that part yet (or you did test it, and it didn't work :P).

I'll admit that it's not something I've tried myself, but I think my first impulse would be to do something similar to what you have (this is also not tested):

# $link_target and $url_just_fetched are known...

my $full_target;

if ( $link_target =~ m{^/} ) {
   # initial "/" means "relative to doc-root":
   my $docroot = $url_just_fetched;
   $docroot =~ s{(?<=[^/]/)[^/].*}{}; # delete everything after first 
+single slash
   $full_target = $docroot . $link_target;
}
elsif ( $link_target !~ /^http/ ) {
   # it's presumably relative to the url just fetched, so:

   if ( $url_just_fetched =~ m{/$} ) {
      $full_target = $url_just_fetched . $link_target;
   }
   # this is the tricky part:
   elsif ( $url_just_fetched =~ /[?&=;:]|\.htm/ )  # probably not a di
+rectory name...
   {
      my $last_slash = rindex( $url_just_fetched, "/" ) + 1;
      $full_target = substr( $url_just_fetched, 0, $last_slash ) . $li
+nk_target;
   }
   else  # assume its a directory name {
      $full_target = join "/", $url_just_fetched, $link_target;
   }
}
# (if $link_target does start with "http", then it's probably complete
+ already)

# last step:

$full_target =~ s/\#.*//;  # in case the link target is a named anchor
+ within a page
[download]

No doubt there'll be some tweaking to do -- especially the part that tries to guess if the last chunk of a url looks like a directory -- but that should get you started. You probably want to make that a separate subroutine (maybe someday it'll be a module).

Thank you, I also have a question about using a hash, what do you mean by use undef for the key. Wouldn't this make all keys the same, and where do I put the url? Thanks for all the help!

I believe the suggestion is that you use the URL for the key, and set the *value* to undef. Then you'll still have a hash with all of the URLs in it as keys.

Like this:

     $fetched{$url} = undef;

Does this make sense?

Linda