andyjack/hash-rename

Intro

Given two directories, hashes each files contents, and renames any matching files in the target dir to the file with the same name in the source dir.

In the following example:

t/123 - contents are "abc"
t/321 - contents are "ABC"
t2/foo - contents are "ABC"
t2/bar - contents are "abc"

hash-rename t/ t2/ would rename t2/foo to t2/321 and t2/bar to t2/123.

There are checks to prevent overwriting file with the same hash.

hash-rename won't recurse into subdirs of the source and target directories.

Use Case

I wrote this script to help fix a problem where I had BROKEN~1.MP4 names on a network device, but correct names on a local device. I could have re-rsync'ed the files over, but this was a fun exercise. And it at least saved the trouble of writing the files over the network a second time. The script still has to read the whole file over the network, so no big savings unless there is some asymmetry in bandwith and/or metering on your network connection that the script can take advantage of.

Setup

A cpanfile is present, so if you have cpanm installed, you can pull down the dependencies easily:

cpanm -l local --installdeps .

And you can use local::lib in perl to make use of the modules in the local dir.