Re: fdupes and hardlinking duplicate files

[Sam Watkins <swatkins@fastmail.fm>]

On Fri, Dec 17, 2004 at 07:30:24AM -0800, hcheney wrote:
> I have a rather large collection of files and there is a great deal of
> duplication in the tree. What I would like to do is replace all of the
> identical files with hardlinks - which I can't seem to do with fdupes.

Here's another way to do it which should work with any filenames (unless they
contain newlines - hopefully you don't have any of those!).

#!/bin/sh
fdupes -r . "$@" | tee /tmp/fdupes.out |
perl -ne '
	chomp;
	if (/./) { push @f, $_; }
	else {
		$k = shift @f;
		for (@f) {
			chmod 0666, $_; unlink $_; link $k, $_;
		}
		@f = ();
	}
'

THIS SCRIPT IS SUPPLIED WITH ABSOLUTELY NO WARRANTY!  WHEN IT LINKS ALL YOUR
FILES TOGETHER INTO ONE BIG MESSY LUMP, DON'T SAY I DIDN'T WARN YOU!

i.e. it should work, but test it first.