Date: Wed, 2 Feb 2011 00:03:37 -0500
From: Austin Clements <amdragon@MIT.EDU>
To: notmuch@notmuchmail.org
Subject: Re: [RFC PATCH v2 0/8] Custom query parser, date search, folder
	search, and more
Message-ID: <20110202050336.GB28537@mit.edu>
References: <1295165458-9573-1-git-send-email-amdragon@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1295165458-9573-1-git-send-email-amdragon@mit.edu>
User-Agent: Mutt/1.5.20 (2009-06-14)
Precedence: list

I rebased the query parser against current master.  It's on the
qparser-3 branch at
  http://awakening.csail.mit.edu/git/notmuch.git

At cworth's request, I've folded the database closing bug fix in to
the appropriate patch.  I also stripped out my implementation of
folder searching, since it obviously conflicts with cworth's [1].  I'm
not planning to resend the patches unless asked because there were no
actual code changes.

[1] I still assert the "correct" folder solution is somewhere between
mine and cworth's: rooted and non-recursive by default (non-recursive
is the only thing that makes sense for Maildir++ folders and it would
be silly for the default to depend on folder type), but leveraging the
flexibility of custom query transforms to support Maildir++ folders
and recursive (and maybe non-rooted) searches as well.  Non-rooted and
recursive searches even map to natural syntaxes that align with
Xapian's existing wildcard syntax.

Quoth myself on Jan 16 at  3:10 am:
> This is version 2 of the custom query parser.  It now supports date
> searches with sane syntax, folder searches (without any additions or
> changes to the database, unlike cworth's recent commit), and "tag:*"
> and "-tag:*" queries for finding tagged and untagged messages.  I used
> these features to guide changes to the original design and to validate
> the approach.  This is still RFC, but it's much less raw now.
> 
> In addition to the new features, the core query parser has a bunch of
> cleanups and changes, including completely redone NEAR and ADJ
> operators that now behave essentially the same as they do in Xapian's
> query parser.  I also split the implementation of these out into a
> separate patch for ease of review.
> 
> There's a notable lack of tests in this current series.  I do have a
> pile of tests for the lexer, parser, and generator, but the
> infrastructure for testing them needs cleanup before I send that out.
> 

-- 
Austin Clements                                      MIT/'06/PhD/CSAIL
amdragon@mit.edu                           http://web.mit.edu/amdragon
       Somewhere in the dream we call reality you will find me,
              searching for the reality we call dreams.