|
From: | Anthony Liguori |
Subject: | Re: [Qemu-devel] [RFC 00/14] MAINTAINERS cleanups - please read |
Date: | Mon, 16 Apr 2012 12:42:34 -0500 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:11.0) Gecko/20120329 Thunderbird/11.0.1 |
On 04/16/2012 12:17 PM, Peter Maydell wrote:
On 16 April 2012 17:03, Andreas Färber<address@hidden> wrote:Am 16.04.2012 18:00, schrieb Anthony Liguori:On 04/16/2012 09:31 AM, Andreas Färber wrote:Hello, Sparked by conversations with Anthony and the discussion on a recent KVM call, I've started overhauling our MAINTAINERS file. Patches 1-5 fix syntax issues. Patch 6 documents our orphaned stable trees, as requested by Anthony. Patch 7 drops the orphaned and by now completely busted darwin-user emulation. Apparently we have unwritten eligibility criteria for new maintainers in terms of qemu-devel participation and patch handling quality, but no formal mechanism to handle replacing lost maintainers. The current practice has become for Anthony to expect people listed in a MAINTAINERS section with S: Maintained to handle patches in that area themselves and to supply a [PULL] request to get those changes into qemu.git. This has the downside that patches falling into an area, where a maintainer is listed but not responding, simply bitrot on the list.I think we ought to aggressively downgrade subsystems if this is really a problem. Unfortunately, it's hard to judge whether this is a problem until someone complains about a specific subsystem.Patches 8-11 therefore propose to upgrade some actively maintained sections to Maintained to formalize the Maintained vs. Odd Fixes semantics:Maintained means PULLs from maintainer expected.Yes. More specifically, if something is Maintained, I would expect the patch to always come in through that specific tree.Odd Fixes means Reviewed-by/Acked-by or committer's gut feeling is sufficient.Yes. Odd Fixes and below means the patch is "fair game" but that the listed M: probably ought to at least be consulted.The current status descriptions seem to be a copy from Linux. Could you address Kevin's comment by proposing a change to the descriptions in our copy?Here's my stab at it: Supported: Someone is actually paid to look after this. This is the same as Maintained (see below) but you might have a greater chance of faster response times.
Let's drop Supported as a status.
Maintained: Someone actually looks after it. The maintainer will have a git subtree for this area and patches are expected to go through it. Bug reports will generally be investigated.
* For something to be marked Maintained, there must be a person on M: and there must be a git tree for the subsystem.
Odd Fixes: It has a maintainer but they don't have time to do much other than throw the odd patch in. Patches are applied directly to master; there is no subtree, and no requirement for an Ack from the maintainer for a patch to be applied. Bug reports without reasonable quality patches attached are likely to go unfixed. Orphan: No current maintainer. Send patches to qemu-devel; persistence may be required to get something reviewed and committed, especially if it's a large change. Obsolete: Old code. Something tagged obsolete generally means it has been replaced by a better system and you should be using that.
I like these descriptions a lot more. Regards, Anthony Liguori
(I dropped the parenthetical about becoming a maintainer of an orphan area; if we want to keep that I think an expanded section about how to demonstrate that you might be capable of the job, how/when to add sections for new code/drivers/boards, etc would be more useful.) -- PMM
[Prev in Thread] | Current Thread | [Next in Thread] |