From mboxrd@z Thu Jan 1 00:00:00 1970 From: Loic Dachary Subject: Re: Pull requests : speed up the reviews Date: Mon, 10 Nov 2014 20:50:55 +0100 Message-ID: <5461171F.8080401@dachary.org> References: <545E53B5.3070600@dachary.org> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="rmeu1MUjC3dHcbQATF7KBqGpOW5tC8FuG" Return-path: Received: from mail2.dachary.org ([91.121.57.175]:34443 "EHLO smtp.dmail.dachary.org" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1750923AbaKJTu6 (ORCPT ); Mon, 10 Nov 2014 14:50:58 -0500 In-Reply-To: Sender: ceph-devel-owner@vger.kernel.org List-ID: To: Gregory Farnum Cc: Ceph Development This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --rmeu1MUjC3dHcbQATF7KBqGpOW5tC8FuG Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Hi Greg, > On Mon, 10 Nov 2014, Gregory Farnum wrote: >> I suppose one way of handling this might be to ask everybody to >> dedicate a small amount of time to reviews (as you suggest), but to >> emphasize PR management as much as doing the actual review. >> >> 1) If you don't have assigned PRs, look at the unassigned PRs from >> oldest to newest, and based on the target area assign it to the person= >> you think most appropriate.after doing basic due diligence for issues >> that you can handle (ie, the feature is something you don't know is >> inappropriate; skimming the code doesn't make you cry; etc) >> >> 2) Manage any reviews assigned to you: >> 2a) if you're the wrong target, assign it to somebody more appropria= te >> 2b) If you're the right target but are trying to get somebody else >> acquainted with the code base, maybe assign it to them >> 2c) assign priorities based on importance and age >> >> 3) review PRs in appropriate order at the level they deserve, and then= >> either merge or assign it back to the author I'll do as you suggest. >> In particular, this clearly assigns responsibility for a PR to an >> individual. It lets people offload as necessary, and gives feedback to= >> contributors quickly even if that feedback is just "we appreciate your= >> submission and it is in somebody's queue". It lets people contribute >> to the PR process even if there isn't something they feel qualified to= >> review directly. And it doesn't emphasize the time spent reviewing so >> much as the act of moving a PR through the stages of merging. (A >> 15-minute timebox might sound great, but trust me we have lots of PRs >> that would never make it into the code base if they were done in >> 15-minute chunks.) I see what you mean. However there are few pull requests of that kind in = the queue, or am I understimating most of them ? >> Obviously this is also not a very formal process, but I think maybe if= >> we start approaching things this way it will help move stuff through >> the pipeline faster. Based on Sage's sudden enthusiasm for assigning >> PR owners I think he might agree...? +1 --=20 Lo=EFc Dachary, Artisan Logiciel Libre --rmeu1MUjC3dHcbQATF7KBqGpOW5tC8FuG Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iEYEARECAAYFAlRhFx8ACgkQ8dLMyEl6F20QKwCgopUWLtg+0sMsPdJE2lb/zme1 QRsAn3HB9H4R4k1xbLwMRo97rF2KT+i3 =JTwO -----END PGP SIGNATURE----- --rmeu1MUjC3dHcbQATF7KBqGpOW5tC8FuG--