Re: [Qemu-devel] [RFC] QEMU Code Audit Team

[Kevin Wolf]

Am 10.01.2012 14:39, schrieb Andreas Färber:
> Am 10.01.2012 14:33, schrieb Kevin Wolf:
>> Am 10.01.2012 14:22, schrieb Anthony Liguori:
>>> On 01/10/2012 06:58 AM, Kevin Wolf wrote:
>>>> Probably we need to attack the reviewing problem first: That I review
>>>> all block patches myself worked well as long as we were two or three
>>>> people in that area, but today it doesn't scale any more without
>>>> lowering the review standards - and I don't want to do that. Maybe we
>>>> should introduce something like "One Reviewed-by buys you two
>>>> Signed-off-bys for your own patches" ;-)
>>>
>>> I think one thing that helps is to make sure for maintainers to include 
>>> Reviewed-bys in commits.  The script I use (below) takes a mbox with the 
>>> full 
>>> thread and folks Reviewed-by/Tested-bys into the original patch spitting 
>>> out an 
>>> mbox with just the patches and tags.
>>>
>>> That way people are getting credit in git for doing reviews.  It's a small 
>>> incentive but every little bit helps.
>>>
>>> http://git.codemonkey.ws/cgit/mbox-filter.git/
>>
>> I usually do that, although manually.
>>
>> Of the 487 patches I have committed, 71 have a Reviewed-by tag in the
>> commit message. Maybe I've missed to include it for some, but that's
>> about the ratio that feels realistic to me.
> 
> If you want an incentive, just put up a rule that every patch needs to
> be reviewed by at least the submaintainer and one person apart from the
> author (i.e., SoB + RB/AB + SoB). If a patch is lacking that additional
> review, the author will ping the list.

And get no response, delaying the patch forever.

I already tried this unintentionally when I had just too little time to
do the 85% myself. I told people that they should find someone else to
review their series if they want to have it committed. Stefan reviewed
some of them, but in most cases, the series ended up being reviewed by
myself some weeks later.

Kevin