Date: Sat, 12 Jul 2014 23:40:13 -0400
From: Austin Clements <amdragon@MIT.EDU>
To: David Bremner <david@tethera.net>
Subject: Re: Add support for specifying tags during "notmuch new"
Message-ID: <20140713034013.GY4660@mit.edu>
References: <1399018555-1994-1-git-send-email-dme@dme.org>
	<20140502151800.GC8365@mit.edu>
	<87a98ev1h2.fsf@maritornes.cs.unb.ca>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <87a98ev1h2.fsf@maritornes.cs.unb.ca>
User-Agent: Mutt/1.5.21 (2010-09-15)
Cc: notmuch@notmuchmail.org
Precedence: list

Quoth David Bremner on Jul 12 at  4:08 pm:
> Austin Clements <amdragon@MIT.EDU> writes:
> 
> > What happens when this script dies in the middle (say, your computer
> > loses power or notmuch tag conflicts with something else on the write
> > lock)?  One advantage of the standard "new" tag approach is that it's
> > easy to write a stateless post-new tagging script that can be killed
> > at any point and restarted.  (You're right that post-new has a
> > concurrency issue, but we should fix that in its own right.)
> 
> Hi David, Hi Austin;
> 
> So, what to do about this patch series? Is it actually a bad idea to
> allow specifying tags for notmuch new? We already allow it via config,
> right, so this just makes things more convenient.

I'm not sure if it's a bad idea or not, but it's materially different
from the tags in the config.  The tags in the config persist and any
call to notmuch new from any source (cron, 'G' in Emacs, etc.) will
pick them up.  Tags specified on the command line don't persist in
this way.