On Fri, Aug 26, 2011 at 3:13 AM, Dustin Sallings <dus...@spy.net> wrote: > > On Aug 25, 2011, at 1:55 PM, Paul Davis wrote: > >> Oh right. Not entirely certain how fixable that's going to be. >> >> On Thu, Aug 25, 2011 at 3:49 PM, Jan Lehnardt <j...@apache.org> wrote: >>> >>> On Aug 24, 2011, at 06:04 , Paul Davis wrote: >>> >>>> What commit-hash thing? >>> >>> Where the commit hashes in current git.a.o are different from the test repo. > > Are the tree hashes expected to be different? (i.e. did we rearrange > the contents as well as the commits?) The more similar the code is, the > easier it is to prove where things have drifted apart. I can perhaps help > with that if someone can point me to a couple of repos and the expected > outcome. :) > > -- > dustin sallings > > > >
Basic story is that I translated the git.apache.org mirror script to Python and made some tweaks (that shouldn't affect hashes) but the hashes are apparently different. Original mirroring script is at [1] and my Python version is at [2]. Feel free to ask questions about either. I haven't dug into this but my fear is that the revisions are calculated based on something non-stable. Whether that's the git-svn version or SVN state I dunno. Hopefully its as silly as a typo though. I'd be interesting to hear about anything you find. [1] https://svn.apache.org/repos/infra/infrastructure/trunk/projects/git/ [2] http://friendpaste.com/13B0LEy9ZQAGOybTWJzJJQ