[Slightly OT] Re: checking which files on a CD are not in a git-annex repo

[Olaf TNSB]

> While that is true, a way to directly "diff" a random non-annex
> directory and an annex would be _very_ handy, though.

I've run into a similar (but different) problem too.  So, +1 on a diff
utility - it would be useful.

My use case is a directory that I don't control that I want to mirror. I
don't like the structure & naming of the original; it is big(ish) so I
don't want to have multiple copies on my machine; and finally I want to
make sure any new files are included into my git-annex is a *proper*
dir/name scheme.

So my (probably too complicated) solution (which I'm going to change
slightly after writing all of this) is to use rsync with "don't delete" &
"ignore existing" (I don't have rsync man page in front of me for the exact
flags) to copy new files into my local directory. I'm not worried about
content change which makes the process a little easier.

The next step is to copy any non-symlinked files from my local directory to
a staging area (i.e. a folder called 'To-Do' ;-) ) in my annex and then
git-annex add them.

Finally, I go through my local "copy" of the remote directory, and symlink
any files to /dev/null.

Now...after writing all that down, I think I'm going to change this by
following Joey's suggestion and have the local copy in the annex, prob
under $ANNEX/mirrors so I don't need to worry about maintaining integrity
of that folder too...it also reduces the need to do the last symlink to
/dev/null.

HTH & happy to send a copy of my script if anyone wants it.

Cheers,

Olaf