Indeed, I've committed a fix, and a small bug fix for the new code just now. The merge incorrectly removed some control logic, which caused more unnecessary checks to happen. I already marked this in the PR, but committed a fix only today. BTW as noted in the Readme, the theoretic false positive probability is <<2^-90, almost non-existent. Best regards, Marc Stevens On 3/13/2017 8:48 PM, Jeff King wrote: > On Mon, Mar 13, 2017 at 07:42:17PM +0000, Dan Shumow wrote: > >> Marc just made a commit this morning fixing problems with the merge. >> Please give the latest in feature/performance a try, as that seems to >> eliminate the problem. > Yeah, b17728507 makes the problem go away for me. Thanks. > > FWIW, I have all sha1s on github.com running through this right now > (actually, the ad744c8b7 version), and logging any false-positives on > the collision detection. Nothing so far, after a few hours. > > -Peff