On Sun, Aug 28, 2011 at 3:45 AM, Dustin Sallings <dus...@spy.net> wrote: > > On Aug 26, 2011, at 6:10 PM, Paul Davis wrote: > >> The mirror: >> >> http://git.apache.org/couchdb.git >> >> The writable repo: >> >> http://git-wip-us.apache.org/repos/asf/couchdb.git >> >> I haven't looked into this at all yet myself. I'll be working on this >> tomorrow and Sunday though so if you have questions I'll be around. > > Yikes, sorry. I keep this list kind of hidden away and forget to look > sometimes. > > I ran a tree convergence report thing between the WIP and the mirror > and I found them to have identical source code even though the commits vary: > > http://public.west.spy.net/tmp/apache-wip-vs-mirror.html > > If you look back through the history far enough, you'll see that WIP > has 5 commits that were missing from mirror, one of them apparently getting > split over two commits and that's where your hashes began to diverge. The > trees quickly did not, though. > > -- > dustin sallings > > > >
Interesting find. I emailed the guy that runs the git mirrors to see if he had any ideas. I'll try re-running the actual mirror scripts later to see if its me that did something weird with the commits or if it's a git-svn difference.