From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:41386) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XgbLV-0002Sb-Mw for qemu-devel@nongnu.org; Tue, 21 Oct 2014 11:27:43 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1XgbLQ-0004Tx-8u for qemu-devel@nongnu.org; Tue, 21 Oct 2014 11:27:37 -0400 Received: from mx1.redhat.com ([209.132.183.28]:54634) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XgbLP-0004Qw-U6 for qemu-devel@nongnu.org; Tue, 21 Oct 2014 11:27:32 -0400 Message-ID: <54467B5B.8000909@redhat.com> Date: Tue, 21 Oct 2014 11:27:23 -0400 From: John Snow MIME-Version: 1.0 References: <20141020102559.GA29278@stefanha-thinkpad.redhat.com> In-Reply-To: Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] Patch checking bot List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Peter Maydell , Stefan Hajnoczi Cc: Fam Zheng , QEMU Developers On 10/20/2014 10:08 AM, Peter Maydell wrote: > On 20 October 2014 11:25, Stefan Hajnoczi wrote: >> Hi, >> At KVM Forum 2014 we discussed a patch checking bot that automates pat= ch >> format checking and smoke testing: >> >> 1. Did the patch submitter include Signed-off-by? >> 2. Does checkpatch.pl pass? >> 3. Does the patch apply to qemu.git/master? >> 4. Does each patch compile? >> 5. Does the series pass make check and qemu-iotests? >> >> Here are some thoughts on the patch checker: >> >> If a patch series passes successfully, no email is sent. If a patch >> series fails, an email with the errors is sent as a reply to the patch >> series email thread. The patch submitter can then respond in case the= re >> are false positive (e.g. from checkpatch.pl) - the bot doesn't care >> about replies but it tells the human reviewers and maintainers what th= e >> patch submitter intends to do. > > Probably also worth having a feature where the cover > letter or patch can have a "patchchecker: no" line in > it to tell the bot to ignore something, so people can > avoid it sending lots of mail for patch series they > know don't apply to mainline (eg ones which depend on > a previous series). > > -- PMM > Maybe it should still check what it can, but squelch the reply to list.=20 Certain automatic checks may still be of value, even if it doesn't apply=20 to master. If we have a website where we can check the bot and patch=20 status, having some output might be nicer than allowing arbitrary skips. Further, maybe the bot could be trained to check the patch series target=20 to see what branch it's supposed to apply to and go from there. If=20 people are good about labeling their stable patches, the bot should be=20 able to check those as well. This might help us tighten up and formalize our subject formatting=20 rules, which would probably also help maintainer work-flow by allowing=20 more robust automation. e.g., "Sorry, you submitted a patch, but I=20 couldn't identify which branch/component you're trying to patch against!" I believe at KVM Forum it was also mentioned that it'd be nice to have=20 the bot reply to patch series where the proper maintainers were missed=20 with a "Hey, I added in who maintains , please include it next time!" --=20 =E2=80=94js