DomainKey-Signature: a=rsa-sha1; c=nofws; d=cs.rpi.edu; h=message-id
	:date:from:mime-version:to:subject:content-type
	:content-transfer-encoding; q=dns; s=default; b=ruz/Egvug30Cfsh9
	eT78f/rqQxP8fE6pN8jETDe5+YAF7+IkzKz+K5lvCN14LEwniH33Kk/fT1e6IoYb
	VBpjLcTgodGwbXJeIpjqh8B5R0Vp2SO0UtmCP/gZCBT+nflwgFhIlLtalPh4Nd7T
	I6A7zfH94YO5wNShEmJTQ4MMjaI=
Message-ID: <4CAB8F24.7060300@cs.rpi.edu>
Date: Tue, 05 Oct 2010 16:48:36 -0400
From: Ethan Glasser-Camp <glasse@cs.rpi.edu>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US;
	rv:1.9.1.12) Gecko/20100915 Thunderbird/3.0.8
MIME-Version: 1.0
To: notmuch@notmuchmail.org
Subject: notmuch synchronization
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: list

I recently got interested in notmuch again and I wanted to see what the
current situation is with regard to synchronization -- specifically
multiple machines each running notmuch, but also notmuch with other
non-notmuch clients. As far as I can tell, this is "possible" but not
easy or clean:

- the message at
http://notmuchmail.org/pipermail/notmuch/2010/002957.html starts a
thread about exactly this subject. This seems like the most promising
approach (using unison to sync mail across all machines, and using
notmuch dump and notmuch restore to propagate the tag database), but I'm
a little scared by the note that:

> Well, only one place where you mark things unread as well with this
> approach. Otherwise updating the database in two places and attempting
> to sync will lead to an unmergable conflict, won't it?

But if I'm understanding things correctly, it should be OK if you mark
unread one place, notmuch dump, restore in the second place, and then
mark something else unread, right? Otherwise you don't have true
disconnected operation, which seems to be to lose tho whole point.

- There's a large number of patches from Michal Sojka, the latest from
April 2010, about an abstract mailstore interface, towards using Git as
an object store; Git has really nice distributed properties, in that
each clone offers fully disconnected operation, and that you can sync
any one against any other one, so something like that sounds pretty
awesome. Unfortunately, Michal wrote that "I do not want this to be
merged yet, but there might be some people interested in testing this.".
Is there any news with these patches? Are they something that could
potentially be merged? At a high level, what kind of behavior do they
implement?

(The most recent thread on this subject was at:
http://notmuchmail.org/pipermail/notmuch/2010/002119.html)

Ethan