[gmx-developers] Gerrit activity
Roland Schulz
roland at utk.edu
Sat Nov 8 22:43:37 CET 2014
Hi,
a few more numbers. Maybe we can encourage more review
by acknowledging each other more for doing review. In that spirit I want to
thank Teemu for being the reviewer of the month October for most review
comments:
select FULL_NAME, count(*) as messages from CHANGE_MESSAGES as m, CHANGES
as c, accounts as a where account_id=AUTHOR_ID and c.CHANGE_ID=m.CHANGE_ID
and AUTHOR_ID<>OWNER_ACCOUNT_ID and month(m.WRITTEN_ON)=10 and
year(m.WRITTEN_ON)=2014 group by FULL_NAME order by messages desc;
FULL_NAME | MESSAGES
--------------------+---------
Jenkins Buildbot | 253
Teemu Murtola | 80
Roland Schulz | 54
Szilárd Páll | 35
Mark Abraham | 35
Berk Hess | 32
Carsten Kutzner | 9
David van der Spoel | 8
Jiri Kraus | 7
Rossen Apostolov | 5
Erik Lindahl | 4
Thomas Ullmann | 4
Christian Wennberg | 1
Viveca Lindahl | 1
John Eblen | 1
Justin Lemkul | 1
(16 rows; 339 ms)
In the last 90 days we did pretty well have having an OK ratio of review
votes to commits created. Largely because Mark did a lot of reviewing for
all of us:
select full_name, reviewed, authored, round(cast(reviewed as
float)/authored,2) as ratio from accounts a left JOIN
(select OWNER_ACCOUNT_ID as ACCOUNT_ID, count(*) as authored from Changes
where status<>'d' and DEST_PROJECT_NAME='gromacs' and now()-created_on<90
group by ACCOUNT_ID) C
ON A.ACCOUNT_ID=C.ACCOUNT_ID left join
(select ACCOUNT_ID, count(distinct c.CHANGE_ID) as reviewed from
PATCH_SET_APPROVALS as P,Changes as C where CATEGORY_ID='Code-Review' and
C.change_id=P.change_id and p.ACCOUNT_ID<>OWNER_ACCOUNT_ID and VALUE!=0 and
DEST_PROJECT_NAME='gromacs' and now()-granted<90 group by ACCOUNT_ID) R
ON R.ACCOUNT_ID=A.ACCOUNT_ID where authored is not null or reviewed is not
null order by ratio, authored;
FULL_NAME | REVIEWED | AUTHORED | RATIO
--------------------+----------+----------+------
Alexey Shvetsov | 1 | NULL | NULL
Paul van Maaren | 1 | NULL | NULL
Thomas Ullmann | 4 | NULL | NULL
Christian Wennberg | 2 | NULL | NULL
John Eblen | 2 | NULL | NULL
Jeroen Engelberts | NULL | 1 | NULL
Åke Sandgren | NULL | 1 | NULL
Jiri Kraus | NULL | 1 | NULL
Christoph Junghans | NULL | 2 | NULL
Rossen Apostolov | 1 | 10 | 0.1
Magnus Lundborg | 3 | 5 | 0.6
Teemu Murtola | 58 | 75 | 0.77
Jan Henning Peters | 1 | 1 | 1.0
Berk Hess | 24 | 22 | 1.09
Roland Schulz | 85 | 65 | 1.31
Erik Lindahl | 21 | 15 | 1.4
Justin Lemkul | 6 | 4 | 1.5
Carsten Kutzner | 7 | 4 | 1.75
Viveca Lindahl | 2 | 1 | 2.0
David van der Spoel | 22 | 9 | 2.44
Mark Abraham | 137 | 40 | 3.43
Szilárd Páll | 26 | 5 | 5.2
(22 rows; 3460 ms)
But the same query for the last 30 days shows that we now all have less
than 2 review votes per change created. That isn't sustainable if we want
to keep our policy of 2 votes per commit.
select full_name, reviewed, authored, round(cast(reviewed as
float)/authored,2) as ratio from accounts a left JOIN
(select OWNER_ACCOUNT_ID as ACCOUNT_ID, count(*) as authored from Changes
where status<>'d' and DEST_PROJECT_NAME='gromacs' and now()-created_on<30
group by ACCOUNT_ID) C
ON A.ACCOUNT_ID=C.ACCOUNT_ID left join
(select ACCOUNT_ID, count(distinct c.CHANGE_ID) as reviewed from
PATCH_SET_APPROVALS as P,Changes as C where CATEGORY_ID='Code-Review' and
C.change_id=P.change_id and p.ACCOUNT_ID<>OWNER_ACCOUNT_ID and VALUE!=0 and
DEST_PROJECT_NAME='gromacs' and now()-granted<30 group by ACCOUNT_ID) R
ON R.ACCOUNT_ID=A.ACCOUNT_ID where authored is not null or reviewed is not
null order by ratio, authored;
FULL_NAME | REVIEWED | AUTHORED | RATIO
--------------------+----------+----------+------
Thomas Ullmann | 4 | NULL | NULL
Christian Wennberg | 1 | NULL | NULL
John Eblen | 2 | NULL | NULL
Jeroen Engelberts | NULL | 1 | NULL
Magnus Lundborg | NULL | 1 | NULL
Jiri Kraus | NULL | 1 | NULL
Teemu Murtola | 13 | 22 | 0.59
Viveca Lindahl | 1 | 1 | 1.0
Szilárd Páll | 1 | 1 | 1.0
Justin Lemkul | 1 | 1 | 1.0
David van der Spoel | 2 | 2 | 1.0
Roland Schulz | 19 | 17 | 1.12
Mark Abraham | 8 | 6 | 1.33
Carsten Kutzner | 3 | 2 | 1.5
Berk Hess | 9 | 6 | 1.5
(15 rows; 3482 ms)
I of course don't suggest that everyone needs to have a ratio over 2 (the
numbers don't capture that some changes are trivial and other are difficult
to review or any other aspects related to size, complexity, and quality).
But we do need the ratio over 2 in average and thus need to do something so
that we all do some more reviews.
Roland
On Sat, Nov 1, 2014 at 1:35 AM, Schulz, Roland <rschulz3 at vols.utk.edu>
wrote:
>
>
> On Sat, Nov 1, 2014 at 12:17 AM, Teemu Murtola <teemu.murtola at gmail.com>
> wrote:
>
>> Interesting graph. I assume that the number of changes is the number of
>> new created changes.
>>
> yes
>
>> It might also be interesting to to include the number of changes that
>> were resolved (merged, abandoned, deleted), and possibly the average
>> in-flight time (from initial upload to resolution) of a change. Not sure
>> how easy the latter is to calculate from available data or to present in
>> the same graph, but these could be useful to get an idea of what one can
>> expect for new changes.
>>
>
> The average number of days it took from a change to be created until it
> was merged:
>
> select avg(m.WRITTEN_ON-c.CREATED_ON),month(m.WRITTEN_ON) as month,
> year(m.WRITTEN_ON) as year from changes as c, change_messages as m where
> c.CHANGE_ID=m.CHANGE_ID and c.STATUS='M' and m.message like 'Change has
> been successfully%' group by month,year order by year,month;
>
> AVG(DATEDIFF('DAY', C.CREATED_ON, M.WRITTEN_ON)) | MONTH | YEAR
> -------------------------------------------------+-------+-----
> 1 | 8 | 2011
> 2 | 9 | 2011
> 14 | 10 | 2011
> 9 | 11 | 2011
> 15 | 12 | 2011
> 6 | 1 | 2012
> 14 | 2 | 2012
> 4 | 3 | 2012
> 3 | 4 | 2012
> 3 | 5 | 2012
> 7 | 6 | 2012
> 10 | 7 | 2012
> 4 | 8 | 2012
> 6 | 9 | 2012
> 11 | 10 | 2012
> 7 | 11 | 2012
> 11 | 12 | 2012
> 7 | 1 | 2013
> 10 | 2 | 2013
> 19 | 3 | 2013
> 28 | 4 | 2013
> 17 | 5 | 2013
> 29 | 6 | 2013
> 18 | 7 | 2013
> 34 | 8 | 2013
> 25 | 9 | 2013
> 28 | 10 | 2013
> 30 | 11 | 2013
> 7 | 12 | 2013
> 21 | 1 | 2014
> 20 | 2 | 2014
> 15 | 3 | 2014
> 6 | 4 | 2014
> 9 | 5 | 2014
> 6 | 6 | 2014
> 9 | 7 | 2014
> 11 | 8 | 2014
> 14 | 9 | 2014
> 14 | 10 | 2014
>
>
>>
>> On Sat, Nov 1, 2014 at 3:33 AM, Roland Schulz <roland at utk.edu> wrote:
>>
>>> I was curious about our monthly activity and made the below graph.
>>> Maybe someone else finds it interesting.
>>> Shows is:
>>> - Number of changes. Includes all projects (e.g. also regressiontests)
>>> and all statuses (e.g. draft)
>>> - Number of patchsets. Also includes all. Shown on right axis.
>>> - Number of comments. Excludes comments by Jenkins or the patch author.
>>> Shown on right
>>>
>>> [image: Inline image 2]
>>> --
>>> ORNL/UT Center for Molecular Biophysics cmb.ornl.gov
>>> 865-241-1537, ORNL PO BOX 2008 MS6309
>>>
>>> --
>>> Gromacs Developers mailing list
>>>
>>> * Please search the archive at
>>> http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List before
>>> posting!
>>>
>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>
>>> * For (un)subscribe requests visit
>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers
>>> or send a mail to gmx-developers-request at gromacs.org.
>>>
>>
>>
>
>
> --
> ORNL/UT Center for Molecular Biophysics cmb.ornl.gov
> 865-241-1537, ORNL PO BOX 2008 MS6309
>
--
ORNL/UT Center for Molecular Biophysics cmb.ornl.gov
865-241-1537, ORNL PO BOX 2008 MS6309
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20141108/6e6ece20/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 28263 bytes
Desc: not available
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20141108/6e6ece20/attachment-0001.png>
More information about the gromacs.org_gmx-developers
mailing list