From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from localhost (localhost [127.0.0.1]) by arlo.cworth.org (Postfix) with ESMTP id AB7C06DE0F82 for ; Sun, 3 Mar 2019 06:56:38 -0800 (PST) X-Virus-Scanned: Debian amavisd-new at cworth.org X-Spam-Flag: NO X-Spam-Score: -0.01 X-Spam-Level: X-Spam-Status: No, score=-0.01 tagged_above=-999 required=5 tests=[AWL=-0.009, SPF_PASS=-0.001] autolearn=disabled Received: from arlo.cworth.org ([127.0.0.1]) by localhost (arlo.cworth.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id gr33WBXGzuSz for ; Sun, 3 Mar 2019 06:56:37 -0800 (PST) Received: from fethera.tethera.net (fethera.tethera.net [198.245.60.197]) by arlo.cworth.org (Postfix) with ESMTPS id 834B96DE0EC6 for ; Sun, 3 Mar 2019 06:56:37 -0800 (PST) Received: from remotemail by fethera.tethera.net with local (Exim 4.89) (envelope-from ) id 1h0SXZ-0008NS-Do for notmuch@notmuchmail.org; Sun, 03 Mar 2019 09:56:33 -0500 Received: (nullmailer pid 1026 invoked by uid 1000); Sun, 03 Mar 2019 14:56:31 -0000 From: David Bremner To: notmuch@notmuchmail.org Subject: Re: WIP2: index user headers In-Reply-To: <20190302154133.25642-1-david@tethera.net> References: <20190302154133.25642-1-david@tethera.net> Date: Sun, 03 Mar 2019 10:56:31 -0400 Message-ID: <87fts4xazk.fsf@tethera.net> MIME-Version: 1.0 Content-Type: text/plain X-BeenThere: notmuch@notmuchmail.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: "Use and development of the notmuch mail system." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 03 Mar 2019 14:56:38 -0000 David Bremner writes: > This obsoletes [1] > This is getting closer to mergable, but it still needs at least to > sanity check the names of user defined prefixes (see point (a) below). > > The main differences from [1] are > > (a) xapian prefixes are no longer defined via upper casing, as this is > locale dependent. The do rely on a ":" separator, hence the need > for some sanitization. > > (b) The caching of user header/prefix information is now done via > string maps, and used more effectively during indexing. I had another thought about user prefixes. I wonder if they should all be forcibly prefixed with something that prevents collisions, to prevent later pain if we add an "official" prefix with the same name. A quick tests suggest it would work to use something like _ so notmuch search --output=files _list:notmuch works. It's a bit ugly, I'll have to play with other options; the main question is whether we think prefixing is needed / worth-it.