[[!comment format=mdwn
 username="arand"
 ip="130.243.226.21"
 subject="comment 7"
 date="2013-08-12T18:07:44Z"
 content="""
Hmm, probably, do you think this way of getting everything and then doing the filtering could work reasonably in git-annex? Assuming \"git cat-file --batch\" is the main bottleneck, reducing the amount of runs for it would still likely be an improvement?
"""]]