[Gluster-devel] Gerrit Statistics - June 2014

[Vijay Bellur]

On 07/05/2014 01:59 AM, Justin Clift wrote:
> On 04/07/2014, at 1:14 PM, Vijay Bellur wrote:
>> Hi All,
>>
>> I have pulled together some statistics from gerrit for fun. The statistics that I have generated are limited by my understanding of gerrit's gsql interface. I don't claim that these stats are 100% accurate - if you notice any aberrations, please let me know :).
>>
>> [1] has number of patches sent by submitters in June 2014, [2] has number of patches sent by submitters in 2014 till June 30th, [3] has number of reviews done by reviewers in June 2014 and [4] has number of reviews done by reviewers in 2014.
>>
>> Cheers,
>> Vijay
>>
>> [1] http://employees.org/~vbellur/glusterfs/stats/patches-201406.txt
>> [2] http://employees.org/~vbellur/glusterfs/stats/patches-2014.txt
>> [3] http://employees.org/~vbellur/glusterfs/stats/reviews-201406.txt
>> [4] http://employees.org/~vbellur/glusterfs/stats/reviews-2014.txt
>
>
> Awesome.  Pranith appears to be a patch generating machine :>, and
> you're absolutely nailing reviews.  With the "reviews" do you reckon
> that's accurate?

Review count includes those reviews that provide a CR vote (-2, -1, +1, 
+2). If different versions of the patchset are reviewed by the same 
reviewer, review count does get incremented as many times. Since there 
has been an effort put in to review multiple times, I think it is worth 
counting that.

>
> Wonder if it's picking up the "Change has been successfully
> cherry-picked as xxx" as messages and stuff as reviews?
>

Only if the CR accompanying that message has a +1 or +2 vote. If a patch 
gets merged without any CR votes, the committer's review count does not 
increment with the query I run. If review comments are passed without a 
CR vote, neither do they get picked up by the query.

-Vijay