[Development] Proposal: Allow contributors to +1 sanity review.

Alan Alpert 416365416c at gmail.com
Wed Aug 14 20:22:28 CEST 2013


On Tue, Aug 13, 2013 at 11:50 PM, Koehne Kai <Kai.Koehne at digia.com> wrote:
>
>> -----Original Message-----
>> From: development-bounces+kai.koehne=digia.com at qt-project.org
>> [mailto:development-bounces+kai.koehne=digia.com at qt-project.org] On
>> Behalf Of Alan Alpert
>> Sent: Tuesday, August 13, 2013 11:18 PM
>> To: André Pönitz
>> Cc: development
>> Subject: Re: [Development] Proposal: Allow contributors to +1 sanity review.
>>
>> [...]
>> To be clear, this discussion is not about altering the functionality of the sanity
>> bot. The current discussion is about the Gerrit interface to the sanity review
>> field. If there were any use for that field other than the bot, then we
>> wouldn't end up having the two conversations so closely tied together.
>
> Regarding this, I think the current setup  is fine. It should be an approvers job to overwrite the bot, not contributors, even if that is inconvenient at times.
>
> The way it's currently implemented, both "Code Review" and "Sanity Review" are sort of orthogonal. I can e.g. give a +2 to a commit that is okay ... except for the sanity issues the bot points out. If the committer however can freely overwrite the bot judgment I'd have to stay away from +2, and then there's actually no point in having a separate 'Sanity Review' field at all ... The Sanity Review bot could as well just give a -1 or +1 in the Code Review.

+2 means that you approve the code going into the project. If there
are known issues and you do not approve it going in, you should be
giving a +1 instead of a +2. Either way (current, or proposed but you
+1), another approver is needed or the contributor needs to resubmit
for your approval. The case where there are no accidental mistakes is
effectively the same.

To my knowledge, there isn't really a point to having a separate
Sanity Review field at all. I don't know why the sanity bot doesn't
just give +/- 1, which would give effectively the same results without
having to maintain an unusual gerrit UI. I have not seen a sanity
review or equivalent in any other gerrit instance.

--
Alan Alpert



More information about the Development mailing list