use Perl

A couple weeks ago, I had an idea as I was walking across the desert. If I could pretend to be base.pm, I could do whatever I wanted whenever a module did a use base, and along the way I could map the @ISA tree. Now, this was just for fun, since I know that every module does not use base.pm.

First, the code:

#!/usr/bin/perl -w
use strict;
 
$INC{'base.pm'} = 1;
 
package base;
 
sub import {
    my $class  = shift;
    my $caller = ( caller )[0];
 
    $main::Isa{$caller} = [ @_ ];
 
    foreach my $package ( @_ )
        {
        eval {  eval "require $package"  };
        }
    }
 
package main;
 
eval {  eval "require $ARGV[0]"  };
 
delete $INC{'base.pm'};
 
display( $ARGV[0], 0 );
 
sub display
    {
    my( $key, $level ) = @_;
 
    my $space = "\t" x ($level + 1);
    print "$space$key\n";
    return unless UNIVERSAL::isa( $main::Isa{$key}, 'ARRAY' );
 
    foreach my $key ( sort { lc $a cmp lc $b } @{ $main::Isa{$key} } )
        {
        display( $key, $level + 1 );
        }
    }

First, I have to make Perl think it has already loaded base.pm, otherwise it will load it when a module tries to use it. The %INC hash keeps track of that, so I create the right key and give it a true value (although you typically find path names there).

Second, I create my base package and its import routine. This is the subroutine that perl calls when I use base qw(foo bar). I do some magic to pull out @ISA, and store it. I do everything with full pacakge specifications because somewhere along the way I was worried about things going kablooey with all my shenanigans (for instance, can I trust vars.pm when I am fooling around like this?)

Third, I go back to the main package, load the module named on the command line, and I am off to the races. Once I am all done, I pretty print the stuff in @Inc.

Fourth, I realize although I have a lot of fun, this is a really stupid thing to do and it is not very illuminating. It is more like an extra credit homework problem than a useful task.