I have mostly been thinking about gcrypt today. This issue needs to be dealt with. The question is, does it really make sense to try to hide the people a git repository is encrypted for? I have posted some thoughts and am coming to the viewpoint that obscuring the identities of users of a repository is not a problem git-annex should try to solve itself, although it also shouldn't get in the way of someone who is able and wants to do that (by using tor, etc).

Finally, I decided to go ahead and add a gcrypt.publish-participants setting to git-remote-gcrypt, and make git-annex set that by default when setting up a gcrypt repository.

Some promising news from the ghc build on arm. I got a working ghc, and even ghci works. Which would make the template haskell in the webapp etc avaialble on arm without the current horrible hacks. Have not managed to build the debian ghc package successfully yet though.

Also, fixed a bug that made git annex sync not pull/push with a local repository that had not yet been initialized for use with git-annex.

Today's work was sponsored by Stanley Yamane.