#!/usr/bin/perl
# dirdiff.pl -- files diff for two directories.
use strict;
use warnings;
use diagnostics;
use File::Find qw(find);
use Getopt::Long;
use Pod::Usage;
use Cwd qw(abs_path);
use Sort::Naturally;

our $VERSION = '0.05';
my @opt_exts;
GetOptions(
    'debug=i' => \( my $debug = 0 ),
    'help|?'  => \( my $opt_help ),
    'man'     => \( my $opt_man ),
    'version' => \( my $opt_version ),
    'aonly'   => \( my $opt_aonly ),
    'bonly'   => \( my $opt_bonly ),
    'inter'   => \( my $opt_inter ),
    'jonly'   => \( my $opt_jonly ),
    'plain'   => \( my $opt_plain ),
    'order'   => \( my $opt_order ),
    'ext=s'   => \@opt_exts,
  )
  or pod2usage(2);

if ($opt_version) {
    print "$0 vrs. $VERSION\n";
    exit;
}
pod2usage(1) if $opt_help;
pod2usage( -verbose => 2 ) if $opt_man;
pod2usage("$0: No paths given.") unless @ARGV;
@opt_exts = split ( /,/, join ( ',', @opt_exts ) ) if @opt_exts;

push ( @ARGV, '.' );
my @dir1 = scandir( $ARGV[0], @opt_exts );
my @dir2 = scandir( $ARGV[1], @opt_exts );

if ( not( $opt_bonly or $opt_jonly ) ) {
    print_diff( '<', abs_path( $ARGV[0] ), diff( \@dir1, \@dir2 ) );
}
if ( not( $opt_aonly or $opt_jonly ) ) {
    print_diff( '>', abs_path( $ARGV[1] ), diff( \@dir2, \@dir1 ) );
}
if ( $opt_inter or $opt_jonly ) {
    print_diff( '=', '', isect( \@dir1, \@dir2 ) );
}

sub print_diff {
    my ( $decor, $dir, @filename ) = @_;
    my $prefix = '';
    $prefix .= "$decor " if $opt_plain;
    $prefix .= "$dir" if defined $dir and length $dir; # awkward :-(
    print map "$prefix$_\n", $opt_order ? nsort(@filename) : @filename;
}

sub scandir {
    my ( $dir, @extensions ) = @_;
    my @dirlist;
    find(
        sub {
            return if /^\.$/ or /^\.\.$/;
            if (@extensions) {
                return unless matched( $_, @extensions );
            }
            push ( @dirlist,
                substr( $File::Find::name, index( $File::Find::name, '/' ) ) );
        },
        $dir
    );
    return @dirlist;
}

sub matched {
    my $pat  = shift;
    my @pats = @_;
    return grep $pat =~ /$_$/, @pats;
}

sub diff {
    my $a = shift;
    my $b = shift;
    my %seen;
    my @aonly;
    @seen{@$b} = ();
    for (@$a) {
        push ( @aonly, $_ ) unless exists $seen{$_};
    }
    return @aonly;
}

sub isect {
    my $a = shift;
    my $b = shift;
    my %union;
    my %isect;
    for ( @$a, @$b ) {
        $union{$_}++ && $isect{$_}++;
    }
    return keys %isect;
}

__END__

=head1 NAME

 dirdiff - files diff for two directories.

=head1 SYNOPSIS

dirdiff [options] path1 [path2]

 Options:

  --debug     set debug level, default is off
  --help      brief help message
  --man       full documentation
  --version   version number
  --aonly     files in a not in b only
  --bonly     files in b not in a only
  --inter     files in a and b
  --jonly     files in a and b only
  --plain     display result with out prefix; '>', '<' or '='
  --order     sorted display of result
  --ext       allowed file extensions

 Switches that don't define a value can be done in long or short form.
 eg:
   dirdiff --man
   dirdiff -m

=head1 OPTIONS

=over 8

=item B<--debug>

Display debug information as program is executed. Control is set by level of the value
passed on the command line. Default value is off (debug == 0).

=item B<--help>

Print a brief help message and exit.

=item B<--man>

Print the manual page (full documentation) and exit.

=item B<--version>

Print the version number and exit.

=item B<--aonly>

Print files in directory a not in directory b.

=item B<--bonly>

Print files in directory b not in directory a.

=item B<--inter>

Print files jointly in both directories (intersection of directory a and directory b).

=item B<--jonly>

Print files jointly in both directories only.

=item B<--plain>

Print results free of decoration, i.e. no '>', '<' or '='.

=item B<--order>

Sort level. Defaults to the sort of no sort. Using this options
will sort the results in ascending order.

=item B<--ext>

Allows for specification of file extension to limit result. May be used multiple times
or with a comma separated list.

=back

=head1 DESCRIPTION

dirdiff is a simple script to show a files diff for two directories. Similar to the
real 'diff', things in directory a are prefixed with a '<'. Things in directory b
are prefixed with a '>' and things in both directories are prefixed with a '='. The
prefix may be removed with the --plain option. Options exist to display results by
themselves or in combination. The default setting is to display files in a, not 
in b, and files in b, not in a. To add the intersection of a and b, use the --inter
option. Note that the default entry for the path2 is '.'.

=head1 AUTHOR

Hugh S. Myers
hsmyers@sdragons.com

=head1 BUGS

None that I know of.

=head1 TODO

 0. Edit POD. ONGOING.
 1. Add some mechanism for wild card specification. DONE.
 2. Add --stats to display date and size information with diff accordingly.
 3. Add --order to allow sorted output. DONE.

=head1 UPDATES

 1. Removed File::Spec::Functions in favor of Cwd, which has the virture of working!
 2. Added refactor ala Aristotle---thank you sir!
 3. Added --ext to allow specifying extensions.
 4. Added --order for sorted output.
 5. Added use Sort::Naturally.
 6. Aristotle refactors his refactor---thanks again!
 7. Restored --order that my implementation of #6. stepped on!

=cut