Here's how it fits in.

On your shell.. Imagine you are user 'bubba'..

mkdir /home/bubba/served
mkdir /etc/a
touch /etc/a/filehere
ln -s /etc/a /home/bubba/served/a

Now for perl..

#!/usr/bin/perl -w
use Cwd;
use strict;

for (qw(
/home/bubba/served
/home/bubba/served/a
/home/bubba/served/a/filehere
)){

    printf "$_ = %s\n",Cwd::abs_path($_);
}
[download]

You get

bubba#mescaline# perl script.pl
/home/bubba/served = /home/bubba/served
/home/bubba/served/a = /etc/a
/home/bubba/served/a/filehere = /etc/a/filehere

I don't know if people using this in the future will be mounting directly to /home/bubba/served , in fact, I'm pretty sure they will not. Things will be mounted somewhere and likely symlinked to /home/bubba/served. ( I believe hard links don't work accross filesystems.. haha).

What the program needs to verify is that whatever the file is, it is 'within' the hierarchy of /home/bubba/served. This is why resolving symlinks won't cut it.