Date: Wed, 4 Apr 2007 07:03:45 -0500 From: Alexander Anderson <a.anderson@utoronto.ca> To: freebsd-questions@freebsd.org Subject: Re: ISO Image Size Increasing Message-ID: <20070404120345.GA95748@upful.org> In-Reply-To: <1d3ed48c0704031200w27431474h46a3f482f65b9bfe@mail.gmail.com> References: <4342.12.170.206.13.1175622392.squirrel@admintool.trueband.net> <1d3ed48c0704031200w27431474h46a3f482f65b9bfe@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
>>The image copied from the CD is approximately 234 MB in size, and the
>>image created by mkisofs is 664 MB.
> It sounds like you may be running into a hardlink issue with iso9660.
Yes, ISO-9660 file system does not assign the same inode to hard links.
I had to write a Perl script that finds identical files and links them
(see below).
Hope it helps.
===begin hardlink.pl===
#!/usr/bin/perl
#
# $Id: hardlink.pl,v 1.2 2007/03/29 01:20:53 alex Exp $
use File::Find;
use strict;
die "Usage: $0 file ...\n" unless @ARGV;
my %count;
my %files;
find({ wanted => \&wanted, no_chdir => 1 }, @ARGV);
sub wanted {
next unless -f;
next if -l;
print "$_\n";
my $md5 = `md5 -q $_`; # shorter than Digest::MD5 (am I lazy)
chomp $md5;
$md5 =~ /^[0-9a-f]{32}$/ or die 'md5 failed';
$count{$md5}++;
push(@{ $files{$md5} }, $_);
}
for my $md5 (grep { $count{$_} > 1 } keys %count) {
my @files = @{ $files{$md5} };
my $source = shift @files;
for my $target (@files) {
system("ln -fv $source $target") == 0 or die;
}
}
__END__
=head1 NAME
hardlink.pl - find copies of files and create hard links instead
=head1 SYNOPSIS
hardlink.pl file ...
=head1 DESCRIPTION
Newsgroups: fa.netbsd.tech.kern
From: Wolfgang Solfrank <w...@tools.de>
Subject: Re: hard links in mounted cd9660 file system
Date: Thu, 3 Mar 2005 13:31:42 GMT
Message-ID: <fa.crmoqrl.l1ik3t@ifi.uio.no>
Hmm, the problem is that there is no good way to know that two files
are hardlinks on a 9660 filesystem. 9660 doesn't have a concept of
inodes as is common in standard unix filesystems. Instead, the
information about the file is stored in the directory entry. This
means that the two directory entries pointing to the same data blocks
may in fact describe two different files (e.g. the may have different
owner or permission, or they may even differ in size!).
Currently, the inode number shown by 9660 is just the offset of the
directory entry of the file relative to the disk/partition, with the
special case for directories, where we use the start of the directory
itself, i.e. the offset of the '.' entry. This way, it's quite easy
to determine the file attributes given the inode number.
=head1 AUTHOR
Alexander Anderson <a.anderson@utoronto.ca>
=cut
===end hardlink.pl===
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070404120345.GA95748>