[[!comment format=mdwn
 username="joey"
 subject="""comment 1"""
 date="2015-09-21T17:09:16Z"
 content="""
Analysis seems to make sense. Note that this happens only when gcrypt is
doing a push, not a pull, so there needs to be a local change made for the
sync to try to push out for the bug to occur. Reproduced by adding a file
and then syncing.

It seems that gcrypt recovers fairly well; after setting the local
remote.foo.gcrypt-id to a new value, when the network comes back, it
re-sets it back to the real value next time a pull is made. Does this
bug result in any problem other than noise in the log?

Anyway, it's certianly a gcrypt bug, and I don't see what git-annex can do
about it. Opened a bug there: <https://github.com/bluss/git-remote-gcrypt/issues/20>

Doesn't seem that this would be too hard to fix in gcrypt..
"""]]