As Shawn already knows, OpenAFS are in the process of moving over to
gerrit for code review. I'll write something up about exactly what
we've done when we finally get there. So far, people seem very
impressed with gerrit, though.
One issue that has been raised (and reported as GERRIT-238) is that
Gerrit seems to use the same email message-ID when initially notifying
a user about a change, and then when subsequently requesting a code
review from an additional reviewer - it looks like both of these
messages are coming from NewChangeSender, which does use a constant
(per change) Message-ID.
Thoughts?
Simon.
Yes, it does that, I forgot about that too.
The rationale here is that when we later need to publish comments,
Gerrit uses that same Message-ID as the In-Reply-To header, so that
the comments thread into the same thread as the original review
request. Since only one email gets sent out to all interested
parties, we generate the same Message-ID in all "please review"
messages so that all comments get nested under it.
But if you really pay attention to the threading, you'll see its only
one level deep. We never thread replies-to-replies, everything
threads as a reply to the original review request. That's partly
because we don't keep track of the reply Message-IDs and thus can't
use them in the future as In-Reply-To headers.
This was opened by Andrew yesterday as
http://jira.source.android.com/jira/browse/GERRIT-238.
I'm willing to fix it... but I'm not sure how to approach it in a way
that preserves this property of being able to at least thread the
reply messages onto the same "review thread".