[[!comment format=mdwn
 username="https://launchpad.net/~arand"
 nickname="arand"
 subject="comment 2"
 date="2013-02-27T15:53:37Z"
 content="""
Hmm, indeed, after further testing it seems like the increased time due to the duplicate branch seems to have been a random quirk, bleh :(

But shouldn't it theoretically be possible to optimize out much of the overhead of multiple very-similar (though not identical) branches though?

I've experimented around with ls-tree and cat-file in a bash script[1], and in this primitive implementation the running time seems to be considerably lower (~0.3s vs ~4s), with much less overhead for extra very-similar branches (~0.7s vs ~37s)

Am I missing some key element that's the reason for the time taken by git annex unused?


[1] primitive annex unused script: [https://gitorious.org/arand-scripts/arand-scripts/blobs/master/annex-funused](https://gitorious.org/arand-scripts/arand-scripts/blobs/master/annex-funused)

timing script: [https://gitorious.org/arand-scripts/arand-scripts/blobs/master/annex-testunused](https://gitorious.org/arand-scripts/arand-scripts/blobs/master/annex-testunused)

"""]]